mradermacher/R1-SGG-7B-i1-GGUF
8B
•
Updated
•
4.61k
Dataset Viewer
image_id
string | image
image | prompt_open
string | prompt_close
string | objects
string | relationships
string |
|---|---|---|---|---|---|
498334
|
Generate a structured scene graph for an image of size (1024 x 768) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 768) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "arm.1", "bbox": [35, 461, 371, 618]}, {"id": "boy.2", "bbox": [30, 31, 457, 767]}, {"id": "food.3", "bbox": [903, 585, 1023, 663]}, {"id": "food.4", "bbox": [809, 603, 930, 661]}, {"id": "hair.5", "bbox": [220, 32, 422, 189]}, {"id": "hand.6", "bbox": [249, 462, 369, 575]}, {"id": "handle.7", "bbox": [278, 446, 395, 559]}, {"id": "plate.8", "bbox": [381, 430, 542, 507]}, {"id": "pole.9", "bbox": [594, 208, 614, 382]}, {"id": "roof.10", "bbox": [79, 64, 229, 99]}, {"id": "shirt.11", "bbox": [67, 197, 414, 767]}, {"id": "sign.12", "bbox": [519, 192, 669, 388]}, {"id": "logo.13", "bbox": [231, 313, 290, 397]}, {"id": "man.14", "bbox": [31, 32, 507, 764]}]
|
[{"subject": "logo.13", "predicate": "on", "object": "shirt.11"}, {"subject": "boy.2", "predicate": "has", "object": "arm.1"}, {"subject": "man.14", "predicate": "has", "object": "hand.6"}, {"subject": "man.14", "predicate": "holding", "object": "plate.8"}, {"subject": "sign.12", "predicate": "with", "object": "pole.9"}]
|
|
498335
|
Generate a structured scene graph for an image of size (662 x 1000) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (662 x 1000) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "railing.1", "bbox": [114, 329, 458, 415]}, {"id": "clock.2", "bbox": [410, 29, 651, 917]}, {"id": "face.3", "bbox": [418, 83, 568, 249]}, {"id": "flower.4", "bbox": [323, 850, 536, 977]}, {"id": "pole.5", "bbox": [86, 55, 139, 98]}, {"id": "pot.6", "bbox": [313, 850, 364, 887]}, {"id": "pot.7", "bbox": [364, 920, 480, 981]}, {"id": "sign.8", "bbox": [0, 12, 475, 335]}, {"id": "street.9", "bbox": [0, 682, 661, 801]}, {"id": "sign.10", "bbox": [0, 23, 430, 291]}]
|
[{"subject": "flower.4", "predicate": "in", "object": "pot.7"}, {"subject": "clock.2", "predicate": "with", "object": "face.3"}]
|
|
498336
|
Generate a structured scene graph for an image of size (768 x 1024) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (768 x 1024) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "light.1", "bbox": [80, 122, 169, 163]}, {"id": "tire.2", "bbox": [364, 413, 629, 801]}, {"id": "window.3", "bbox": [110, 185, 143, 251]}, {"id": "truck.4", "bbox": [28, 95, 653, 786]}, {"id": "truck.5", "bbox": [22, 50, 739, 1012]}, {"id": "man.6", "bbox": [162, 98, 370, 486]}, {"id": "car.7", "bbox": [63, 158, 765, 804]}]
|
[{"subject": "light.1", "predicate": "on", "object": "truck.4"}, {"subject": "tire.2", "predicate": "on", "object": "truck.5"}, {"subject": "man.6", "predicate": "on", "object": "truck.5"}, {"subject": "window.3", "predicate": "on", "object": "truck.5"}]
|
|
498337
|
Generate a structured scene graph for an image of size (467 x 276) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (467 x 276) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "building.1", "bbox": [226, 100, 368, 192]}, {"id": "car.2", "bbox": [26, 171, 93, 213]}, {"id": "horse.3", "bbox": [330, 170, 395, 243]}, {"id": "tire.4", "bbox": [146, 175, 165, 194]}, {"id": "vehicle.5", "bbox": [124, 164, 177, 208]}, {"id": "car.6", "bbox": [123, 157, 180, 211]}]
|
[{"subject": "horse.3", "predicate": "near", "object": "building.1"}, {"subject": "building.1", "predicate": "near", "object": "horse.3"}]
|
|
498338
|
Generate a structured scene graph for an image of size (1024 x 682) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 682) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "guy.1", "bbox": [502, 48, 960, 679]}, {"id": "person.2", "bbox": [3, 104, 534, 659]}, {"id": "shirt.3", "bbox": [519, 277, 788, 680]}]
|
[{"subject": "shirt.3", "predicate": "on", "object": "guy.1"}]
|
|
498339
|
Generate a structured scene graph for an image of size (1024 x 768) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 768) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "bag.1", "bbox": [706, 580, 783, 766]}, {"id": "chair.2", "bbox": [737, 566, 929, 764]}, {"id": "man.3", "bbox": [452, 428, 651, 753]}, {"id": "man.4", "bbox": [665, 373, 807, 671]}, {"id": "woman.5", "bbox": [633, 408, 920, 764]}, {"id": "girl.6", "bbox": [381, 368, 494, 656]}, {"id": "glass.7", "bbox": [668, 481, 695, 551]}, {"id": "hair.8", "bbox": [240, 376, 307, 482]}, {"id": "hair.9", "bbox": [414, 367, 493, 491]}, {"id": "jacket.10", "bbox": [455, 495, 652, 696]}, {"id": "jacket.11", "bbox": [494, 407, 637, 508]}, {"id": "pant.12", "bbox": [95, 635, 231, 694]}, {"id": "pant.13", "bbox": [239, 650, 461, 766]}, {"id": "pant.14", "bbox": [635, 637, 837, 767]}, {"id": "shirt.15", "bbox": [701, 414, 791, 494]}, {"id": "shirt.16", "bbox": [227, 473, 440, 709]}, {"id": "shirt.17", "bbox": [18, 492, 201, 690]}, {"id": "table.18", "bbox": [169, 485, 993, 767]}, {"id": "woman.19", "bbox": [233, 376, 311, 490]}, {"id": "woman.20", "bbox": [13, 409, 241, 743]}, {"id": "man.21", "bbox": [214, 413, 472, 765]}, {"id": "woman.22", "bbox": [632, 678, 816, 764]}, {"id": "woman.23", "bbox": [369, 358, 502, 656]}, {"id": "man.24", "bbox": [502, 365, 639, 522]}, {"id": "vase.25", "bbox": [328, 330, 364, 400]}]
|
[{"subject": "man.21", "predicate": "wearing", "object": "pant.13"}, {"subject": "man.21", "predicate": "wearing", "object": "shirt.16"}, {"subject": "man.3", "predicate": "wearing", "object": "jacket.10"}, {"subject": "woman.22", "predicate": "wearing", "object": "pant.14"}, {"subject": "woman.23", "predicate": "with", "object": "hair.9"}, {"subject": "woman.20", "predicate": "wearing", "object": "pant.12"}, {"subject": "woman.20", "predicate": "wearing", "object": "shirt.17"}, {"subject": "woman.19", "predicate": "with", "object": "hair.8"}, {"subject": "man.4", "predicate": "wearing", "object": "shirt.15"}, {"subject": "man.24", "predicate": "wearing", "object": "jacket.11"}, {"subject": "bag.1", "predicate": "hanging from", "object": "chair.2"}, {"subject": "glass.7", "predicate": "above", "object": "table.18"}]
|
|
498340
|
Generate a structured scene graph for an image of size (1024 x 680) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 680) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "door.1", "bbox": [654, 242, 782, 426]}, {"id": "person.2", "bbox": [839, 296, 866, 369]}, {"id": "tire.3", "bbox": [668, 241, 777, 423]}, {"id": "wheel.4", "bbox": [478, 362, 647, 602]}, {"id": "wheel.5", "bbox": [196, 421, 350, 531]}, {"id": "window.6", "bbox": [879, 210, 1018, 268]}, {"id": "wheel.7", "bbox": [660, 220, 781, 426]}, {"id": "sign.8", "bbox": [209, 242, 279, 342]}]
|
[{"subject": "tire.3", "predicate": "on", "object": "door.1"}]
|
|
498341
|
Generate a structured scene graph for an image of size (767 x 1024) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (767 x 1024) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "arm.1", "bbox": [464, 352, 540, 483]}, {"id": "arm.2", "bbox": [245, 204, 358, 425]}, {"id": "boot.3", "bbox": [314, 849, 523, 934]}, {"id": "hand.4", "bbox": [275, 408, 360, 523]}, {"id": "hand.5", "bbox": [513, 466, 567, 545]}, {"id": "head.6", "bbox": [373, 41, 492, 180]}, {"id": "jacket.7", "bbox": [242, 141, 543, 508]}, {"id": "leg.8", "bbox": [401, 483, 477, 855]}, {"id": "leg.9", "bbox": [258, 453, 441, 875]}, {"id": "pant.10", "bbox": [255, 450, 474, 866]}, {"id": "person.11", "bbox": [252, 40, 572, 946]}, {"id": "ski.12", "bbox": [290, 861, 568, 960]}]
|
[{"subject": "leg.9", "predicate": "of", "object": "person.11"}, {"subject": "hand.4", "predicate": "of", "object": "person.11"}, {"subject": "hand.5", "predicate": "of", "object": "person.11"}, {"subject": "arm.2", "predicate": "of", "object": "person.11"}, {"subject": "arm.1", "predicate": "of", "object": "person.11"}, {"subject": "head.6", "predicate": "of", "object": "person.11"}]
|
|
498342
|
Generate a structured scene graph for an image of size (1024 x 768) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 768) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "boat.1", "bbox": [695, 122, 1003, 219]}, {"id": "flag.2", "bbox": [281, 322, 371, 426]}, {"id": "girl.3", "bbox": [367, 296, 463, 472]}, {"id": "leaf.4", "bbox": [407, 0, 1022, 174]}, {"id": "shirt.5", "bbox": [374, 323, 432, 387]}, {"id": "tree.6", "bbox": [423, 0, 964, 221]}]
|
[{"subject": "girl.3", "predicate": "holding", "object": "flag.2"}, {"subject": "leaf.4", "predicate": "on", "object": "tree.6"}]
|
|
498345
|
Generate a structured scene graph for an image of size (1024 x 685) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 685) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "basket.1", "bbox": [598, 483, 697, 546]}, {"id": "bottle.2", "bbox": [943, 322, 968, 439]}, {"id": "table.3", "bbox": [499, 349, 1020, 679]}, {"id": "coat.4", "bbox": [172, 188, 382, 484]}, {"id": "coat.5", "bbox": [16, 221, 276, 584]}, {"id": "cup.6", "bbox": [779, 407, 856, 498]}, {"id": "glass.7", "bbox": [419, 296, 458, 366]}, {"id": "glass.8", "bbox": [194, 349, 245, 415]}, {"id": "glass.9", "bbox": [654, 315, 684, 378]}, {"id": "glass.10", "bbox": [309, 267, 339, 336]}, {"id": "man.11", "bbox": [371, 169, 539, 679]}, {"id": "man.12", "bbox": [173, 126, 383, 680]}, {"id": "person.13", "bbox": [568, 165, 766, 507]}, {"id": "person.14", "bbox": [16, 138, 271, 679]}, {"id": "jacket.15", "bbox": [895, 243, 993, 349]}, {"id": "jacket.16", "bbox": [372, 239, 543, 493]}, {"id": "kid.17", "bbox": [887, 271, 970, 380]}, {"id": "men.18", "bbox": [172, 128, 542, 684]}, {"id": "pant.19", "bbox": [90, 476, 240, 684]}, {"id": "people.20", "bbox": [15, 126, 1023, 683]}, {"id": "person.21", "bbox": [969, 190, 1023, 314]}, {"id": "person.22", "bbox": [895, 194, 993, 350]}, {"id": "woman.23", "bbox": [775, 186, 913, 408]}, {"id": "basket.24", "bbox": [580, 474, 681, 578]}, {"id": "woman.25", "bbox": [51, 118, 218, 518]}, {"id": "woman.26", "bbox": [21, 140, 302, 680]}, {"id": "man.27", "bbox": [161, 128, 408, 669]}, {"id": "woman.28", "bbox": [559, 156, 765, 507]}, {"id": "woman.29", "bbox": [886, 195, 989, 338]}]
|
[{"subject": "basket.24", "predicate": "on", "object": "table.3"}, {"subject": "woman.25", "predicate": "holding", "object": "glass.8"}, {"subject": "bottle.2", "predicate": "on", "object": "table.3"}, {"subject": "woman.26", "predicate": "holding", "object": "glass.8"}, {"subject": "man.11", "predicate": "wearing", "object": "jacket.16"}, {"subject": "man.27", "predicate": "wearing", "object": "coat.4"}, {"subject": "person.14", "predicate": "wearing", "object": "pant.19"}, {"subject": "person.14", "predicate": "wearing", "object": "coat.5"}, {"subject": "person.14", "predicate": "holding", "object": "glass.8"}, {"subject": "man.12", "predicate": "holding", "object": "glass.10"}, {"subject": "man.11", "predicate": "holding", "object": "glass.7"}, {"subject": "person.13", "predicate": "holding", "object": "glass.9"}, {"subject": "woman.29", "predicate": "wearing", "object": "jacket.15"}]
|
|
498346
|
Generate a structured scene graph for an image of size (1024 x 678) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 678) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "arm.1", "bbox": [313, 341, 356, 451]}, {"id": "bike.2", "bbox": [275, 455, 483, 677]}, {"id": "building.3", "bbox": [174, 0, 410, 215]}, {"id": "bus.4", "bbox": [149, 212, 456, 355]}, {"id": "hat.5", "bbox": [585, 293, 623, 331]}, {"id": "helmet.6", "bbox": [363, 285, 420, 314]}, {"id": "man.7", "bbox": [291, 284, 474, 619]}, {"id": "tree.8", "bbox": [773, 0, 1023, 210]}, {"id": "wheel.9", "bbox": [371, 531, 480, 677]}]
|
[{"subject": "man.7", "predicate": "on", "object": "bike.2"}, {"subject": "wheel.9", "predicate": "on", "object": "bike.2"}, {"subject": "bike.2", "predicate": "with", "object": "man.7"}, {"subject": "man.7", "predicate": "with", "object": "arm.1"}, {"subject": "arm.1", "predicate": "on", "object": "man.7"}]
|
|
498348
|
Generate a structured scene graph for an image of size (768 x 1024) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (768 x 1024) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "bird.1", "bbox": [420, 347, 468, 391]}, {"id": "car.2", "bbox": [23, 375, 649, 978]}]
|
[{"subject": "bird.1", "predicate": "on", "object": "car.2"}]
|
|
498349
|
Generate a structured scene graph for an image of size (1024 x 768) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 768) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "snow.1", "bbox": [0, 462, 1022, 763]}, {"id": "hill.2", "bbox": [0, 155, 355, 527]}, {"id": "jacket.3", "bbox": [635, 533, 696, 604]}, {"id": "jacket.4", "bbox": [136, 668, 200, 760]}, {"id": "man.5", "bbox": [631, 509, 698, 676]}, {"id": "mountain.6", "bbox": [207, 305, 762, 558]}, {"id": "pant.7", "bbox": [641, 598, 692, 668]}, {"id": "people.8", "bbox": [138, 633, 201, 765]}, {"id": "skier.9", "bbox": [36, 490, 801, 766]}, {"id": "ski.10", "bbox": [619, 637, 710, 690]}, {"id": "woman.11", "bbox": [125, 627, 206, 764]}, {"id": "person.12", "bbox": [628, 512, 712, 685]}, {"id": "person.13", "bbox": [124, 633, 198, 764]}]
|
[{"subject": "woman.11", "predicate": "wearing", "object": "jacket.4"}, {"subject": "person.12", "predicate": "wearing", "object": "ski.10"}, {"subject": "person.12", "predicate": "wearing", "object": "jacket.3"}, {"subject": "man.5", "predicate": "in", "object": "mountain.6"}, {"subject": "person.12", "predicate": "on", "object": "snow.1"}, {"subject": "person.13", "predicate": "on", "object": "snow.1"}, {"subject": "skier.9", "predicate": "wearing", "object": "jacket.3"}, {"subject": "skier.9", "predicate": "wearing", "object": "jacket.4"}, {"subject": "man.5", "predicate": "wearing", "object": "pant.7"}]
|
|
498350
|
Generate a structured scene graph for an image of size (765 x 1024) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (765 x 1024) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "child.1", "bbox": [652, 603, 761, 1020]}, {"id": "child.2", "bbox": [264, 490, 473, 1019]}, {"id": "child.3", "bbox": [45, 480, 267, 949]}, {"id": "coat.4", "bbox": [704, 294, 764, 431]}, {"id": "kid.5", "bbox": [0, 350, 764, 1022]}, {"id": "man.6", "bbox": [682, 245, 764, 614]}, {"id": "man.7", "bbox": [535, 222, 614, 471]}]
|
[{"subject": "child.2", "predicate": "standing on", "object": "kid.5"}, {"subject": "child.3", "predicate": "standing on", "object": "kid.5"}, {"subject": "child.1", "predicate": "standing on", "object": "kid.5"}, {"subject": "man.6", "predicate": "wearing", "object": "coat.4"}]
|
|
498352
|
Generate a structured scene graph for an image of size (1024 x 770) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 770) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "building.1", "bbox": [895, 132, 1021, 254]}, {"id": "building.2", "bbox": [0, 0, 206, 227]}, {"id": "window.3", "bbox": [99, 96, 146, 168]}, {"id": "window.4", "bbox": [104, 108, 174, 178]}]
|
[{"subject": "window.4", "predicate": "on", "object": "building.2"}, {"subject": "window.3", "predicate": "on", "object": "building.2"}]
|
|
498353
|
Generate a structured scene graph for an image of size (1024 x 768) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 768) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "curtain.1", "bbox": [264, 290, 366, 465]}, {"id": "train.2", "bbox": [289, 0, 1022, 608]}, {"id": "window.3", "bbox": [619, 161, 662, 287]}, {"id": "window.4", "bbox": [660, 154, 708, 299]}, {"id": "window.5", "bbox": [712, 141, 816, 327]}, {"id": "window.6", "bbox": [529, 158, 556, 263]}, {"id": "window.7", "bbox": [546, 157, 627, 285]}, {"id": "window.8", "bbox": [464, 144, 549, 274]}, {"id": "window.9", "bbox": [376, 162, 409, 223]}, {"id": "vehicle.10", "bbox": [395, 105, 753, 341]}, {"id": "window.11", "bbox": [451, 174, 508, 235]}, {"id": "window.12", "bbox": [584, 174, 731, 272]}]
|
[{"subject": "window.9", "predicate": "of", "object": "train.2"}, {"subject": "window.8", "predicate": "on", "object": "vehicle.10"}, {"subject": "window.11", "predicate": "on", "object": "vehicle.10"}, {"subject": "window.5", "predicate": "on", "object": "train.2"}, {"subject": "window.4", "predicate": "on", "object": "train.2"}, {"subject": "window.12", "predicate": "on", "object": "train.2"}, {"subject": "window.7", "predicate": "on", "object": "train.2"}, {"subject": "window.3", "predicate": "on", "object": "train.2"}, {"subject": "window.6", "predicate": "on", "object": "train.2"}]
|
|
498354
|
Generate a structured scene graph for an image of size (819 x 1024) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (819 x 1024) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "man.1", "bbox": [22, 52, 797, 881]}, {"id": "man.2", "bbox": [460, 33, 730, 681]}, {"id": "fence.3", "bbox": [0, 0, 819, 496]}, {"id": "hair.4", "bbox": [547, 37, 628, 96]}, {"id": "hair.5", "bbox": [96, 55, 225, 148]}, {"id": "hand.6", "bbox": [105, 466, 167, 540]}, {"id": "pole.7", "bbox": [2, 7, 286, 58]}, {"id": "pole.8", "bbox": [748, 0, 765, 155]}, {"id": "pole.9", "bbox": [614, 52, 776, 76]}, {"id": "racket.10", "bbox": [529, 212, 789, 332]}, {"id": "racket.11", "bbox": [61, 364, 160, 540]}, {"id": "shirt.12", "bbox": [190, 135, 536, 431]}, {"id": "shirt.13", "bbox": [465, 137, 709, 384]}, {"id": "shoe.14", "bbox": [712, 711, 800, 879]}, {"id": "shoe.15", "bbox": [28, 787, 194, 870]}, {"id": "short.16", "bbox": [235, 375, 580, 658]}, {"id": "short.17", "bbox": [523, 356, 655, 524]}, {"id": "sock.18", "bbox": [124, 794, 192, 834]}, {"id": "man.19", "bbox": [26, 202, 386, 610]}, {"id": "person.20", "bbox": [118, 177, 676, 759]}, {"id": "shoe.21", "bbox": [659, 723, 776, 877]}, {"id": "person.22", "bbox": [248, 205, 658, 818]}, {"id": "person.23", "bbox": [549, 153, 668, 642]}]
|
[{"subject": "man.1", "predicate": "has", "object": "hair.5"}, {"subject": "man.1", "predicate": "holding", "object": "racket.11"}, {"subject": "man.2", "predicate": "holding", "object": "racket.10"}, {"subject": "man.1", "predicate": "wearing", "object": "short.16"}, {"subject": "man.2", "predicate": "wearing", "object": "short.17"}, {"subject": "man.2", "predicate": "wearing", "object": "shirt.13"}, {"subject": "man.1", "predicate": "wearing", "object": "shirt.12"}, {"subject": "shoe.14", "predicate": "on", "object": "man.1"}, {"subject": "pole.7", "predicate": "on", "object": "fence.3"}, {"subject": "pole.8", "predicate": "on", "object": "fence.3"}, {"subject": "pole.9", "predicate": "on", "object": "fence.3"}]
|
|
498355
|
Generate a structured scene graph for an image of size (683 x 1024) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (683 x 1024) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "box.1", "bbox": [0, 413, 683, 1023]}, {"id": "ear.2", "bbox": [450, 217, 499, 283]}, {"id": "face.3", "bbox": [275, 162, 455, 381]}, {"id": "fork.4", "bbox": [0, 945, 320, 1018]}, {"id": "fork.5", "bbox": [533, 331, 591, 516]}, {"id": "woman.6", "bbox": [143, 3, 627, 588]}, {"id": "glass.7", "bbox": [266, 184, 544, 295]}, {"id": "hair.8", "bbox": [266, 27, 547, 556]}, {"id": "hand.9", "bbox": [490, 450, 609, 580]}, {"id": "hand.10", "bbox": [165, 461, 258, 574]}, {"id": "handle.11", "bbox": [0, 951, 139, 993]}, {"id": "nose.12", "bbox": [317, 233, 366, 291]}, {"id": "pizza.13", "bbox": [37, 598, 683, 1024]}, {"id": "tile.14", "bbox": [0, 371, 30, 457]}, {"id": "tile.15", "bbox": [0, 285, 25, 374]}, {"id": "fork.16", "bbox": [544, 327, 596, 454]}]
|
[{"subject": "ear.2", "predicate": "on", "object": "woman.6"}, {"subject": "fork.4", "predicate": "on", "object": "pizza.13"}, {"subject": "glass.7", "predicate": "on", "object": "face.3"}, {"subject": "pizza.13", "predicate": "in", "object": "box.1"}, {"subject": "hand.9", "predicate": "on", "object": "woman.6"}, {"subject": "fork.5", "predicate": "in", "object": "hand.9"}, {"subject": "face.3", "predicate": "on", "object": "woman.6"}, {"subject": "nose.12", "predicate": "on", "object": "face.3"}, {"subject": "fork.16", "predicate": "on", "object": "hand.9"}, {"subject": "fork.4", "predicate": "has", "object": "handle.11"}, {"subject": "woman.6", "predicate": "in", "object": "glass.7"}, {"subject": "woman.6", "predicate": "with", "object": "hair.8"}]
|
|
498356
|
Generate a structured scene graph for an image of size (1024 x 768) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 768) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "building.1", "bbox": [0, 0, 1022, 764]}, {"id": "face.2", "bbox": [373, 274, 544, 432]}, {"id": "roof.3", "bbox": [606, 132, 686, 175]}, {"id": "tower.4", "bbox": [709, 0, 1021, 766]}, {"id": "window.5", "bbox": [709, 524, 737, 617]}, {"id": "window.6", "bbox": [149, 511, 757, 613]}, {"id": "window.7", "bbox": [508, 489, 573, 636]}, {"id": "window.8", "bbox": [640, 524, 675, 614]}, {"id": "window.9", "bbox": [160, 517, 188, 604]}, {"id": "window.10", "bbox": [236, 508, 269, 605]}, {"id": "window.11", "bbox": [892, 508, 998, 669]}, {"id": "window.12", "bbox": [450, 524, 477, 611]}, {"id": "window.13", "bbox": [382, 519, 413, 608]}, {"id": "window.14", "bbox": [0, 246, 107, 470]}, {"id": "window.15", "bbox": [786, 300, 827, 418]}, {"id": "window.16", "bbox": [822, 517, 865, 630]}, {"id": "window.17", "bbox": [955, 240, 1023, 364]}, {"id": "window.18", "bbox": [227, 507, 286, 605]}, {"id": "window.19", "bbox": [377, 513, 403, 606]}, {"id": "window.20", "bbox": [653, 540, 677, 612]}, {"id": "window.21", "bbox": [382, 513, 409, 602]}, {"id": "window.22", "bbox": [447, 517, 478, 608]}]
|
[{"subject": "tower.4", "predicate": "with", "object": "window.11"}, {"subject": "window.7", "predicate": "on", "object": "building.1"}, {"subject": "tower.4", "predicate": "with", "object": "window.16"}, {"subject": "tower.4", "predicate": "with", "object": "window.17"}, {"subject": "tower.4", "predicate": "with", "object": "window.15"}, {"subject": "face.2", "predicate": "on", "object": "building.1"}, {"subject": "window.18", "predicate": "in", "object": "building.1"}, {"subject": "window.5", "predicate": "on", "object": "building.1"}, {"subject": "window.8", "predicate": "on", "object": "building.1"}, {"subject": "window.13", "predicate": "on", "object": "building.1"}, {"subject": "window.10", "predicate": "on", "object": "building.1"}, {"subject": "window.9", "predicate": "on", "object": "building.1"}, {"subject": "window.19", "predicate": "on", "object": "building.1"}, {"subject": "window.12", "predicate": "on", "object": "building.1"}, {"subject": "window.20", "predicate": "on", "object": "building.1"}, {"subject": "window.6", "predicate": "on", "object": "building.1"}, {"subject": "window.21", "predicate": "on", "object": "building.1"}, {"subject": "window.22", "predicate": "on", "object": "building.1"}]
|
|
498358
|
Generate a structured scene graph for an image of size (1024 x 767) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 767) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "board.1", "bbox": [832, 276, 920, 389]}, {"id": "chair.2", "bbox": [324, 582, 444, 765]}, {"id": "chair.3", "bbox": [140, 627, 335, 764]}, {"id": "chair.4", "bbox": [0, 628, 143, 765]}, {"id": "chair.5", "bbox": [201, 518, 314, 713]}, {"id": "chair.6", "bbox": [81, 518, 190, 716]}, {"id": "chair.7", "bbox": [827, 436, 920, 562]}, {"id": "curtain.8", "bbox": [1, 229, 99, 568]}, {"id": "head.9", "bbox": [486, 239, 565, 345]}, {"id": "jacket.10", "bbox": [409, 324, 648, 644]}, {"id": "man.11", "bbox": [411, 240, 650, 766]}, {"id": "man.12", "bbox": [764, 363, 829, 526]}, {"id": "pant.13", "bbox": [435, 557, 615, 764]}, {"id": "plant.14", "bbox": [294, 368, 348, 426]}, {"id": "shirt.15", "bbox": [805, 418, 888, 498]}, {"id": "table.16", "bbox": [0, 557, 380, 764]}, {"id": "table.17", "bbox": [787, 612, 1022, 765]}, {"id": "table.18", "bbox": [904, 425, 969, 528]}, {"id": "tie.19", "bbox": [495, 349, 545, 559]}, {"id": "window.20", "bbox": [281, 278, 337, 418]}, {"id": "woman.21", "bbox": [776, 384, 889, 557]}]
|
[{"subject": "head.9", "predicate": "on", "object": "man.11"}, {"subject": "man.11", "predicate": "with", "object": "head.9"}, {"subject": "man.11", "predicate": "wearing", "object": "tie.19"}, {"subject": "man.11", "predicate": "wearing", "object": "pant.13"}, {"subject": "chair.2", "predicate": "near", "object": "table.16"}, {"subject": "chair.3", "predicate": "near", "object": "table.16"}, {"subject": "chair.4", "predicate": "near", "object": "table.16"}, {"subject": "chair.5", "predicate": "near", "object": "table.16"}, {"subject": "chair.6", "predicate": "near", "object": "table.16"}, {"subject": "woman.21", "predicate": "wearing", "object": "shirt.15"}, {"subject": "woman.21", "predicate": "sitting on", "object": "chair.7"}, {"subject": "man.11", "predicate": "wearing", "object": "jacket.10"}, {"subject": "man.12", "predicate": "near", "object": "woman.21"}]
|
|
498359
|
Generate a structured scene graph for an image of size (799 x 676) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (799 x 676) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "building.1", "bbox": [0, 33, 713, 674]}, {"id": "building.2", "bbox": [381, 28, 762, 606]}, {"id": "building.3", "bbox": [0, 39, 460, 674]}, {"id": "building.4", "bbox": [0, 36, 333, 358]}, {"id": "cap.5", "bbox": [617, 614, 664, 660]}, {"id": "head.6", "bbox": [615, 612, 671, 671]}, {"id": "jacket.7", "bbox": [685, 626, 782, 674]}, {"id": "leaf.8", "bbox": [0, 161, 493, 496]}, {"id": "light.9", "bbox": [302, 541, 344, 610]}, {"id": "light.10", "bbox": [165, 513, 241, 605]}, {"id": "light.11", "bbox": [195, 512, 241, 597]}, {"id": "light.12", "bbox": [46, 512, 81, 577]}, {"id": "man.13", "bbox": [571, 622, 620, 674]}, {"id": "person.14", "bbox": [688, 589, 782, 673]}, {"id": "roof.15", "bbox": [0, 38, 681, 582]}, {"id": "sign.16", "bbox": [310, 394, 461, 557]}, {"id": "clock.17", "bbox": [516, 297, 562, 339]}, {"id": "roof.18", "bbox": [113, 210, 235, 312]}, {"id": "light.19", "bbox": [39, 515, 88, 596]}, {"id": "light.20", "bbox": [188, 529, 251, 603]}]
|
[{"subject": "light.19", "predicate": "on", "object": "building.1"}, {"subject": "light.10", "predicate": "on", "object": "building.1"}, {"subject": "light.11", "predicate": "on", "object": "building.1"}, {"subject": "light.20", "predicate": "on", "object": "building.3"}, {"subject": "light.9", "predicate": "on", "object": "building.3"}, {"subject": "clock.17", "predicate": "on", "object": "building.2"}, {"subject": "person.14", "predicate": "wearing", "object": "jacket.7"}]
|
|
498360
|
Generate a structured scene graph for an image of size (800 x 534) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (800 x 534) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "glove.1", "bbox": [407, 260, 518, 371]}, {"id": "helmet.2", "bbox": [407, 207, 453, 257]}, {"id": "man.3", "bbox": [390, 206, 728, 373]}, {"id": "snow.4", "bbox": [0, 0, 796, 529]}, {"id": "pole.5", "bbox": [265, 54, 417, 389]}, {"id": "ski.6", "bbox": [504, 318, 682, 373]}, {"id": "skier.7", "bbox": [393, 200, 684, 364]}]
|
[{"subject": "man.3", "predicate": "wearing", "object": "glove.1"}, {"subject": "man.3", "predicate": "wearing", "object": "helmet.2"}, {"subject": "skier.7", "predicate": "wearing", "object": "helmet.2"}, {"subject": "skier.7", "predicate": "wearing", "object": "glove.1"}, {"subject": "skier.7", "predicate": "wearing", "object": "ski.6"}]
|
|
498361
|
Generate a structured scene graph for an image of size (1024 x 768) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 768) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "door.1", "bbox": [316, 272, 392, 469]}, {"id": "door.2", "bbox": [779, 287, 822, 485]}, {"id": "train.3", "bbox": [3, 203, 587, 571]}, {"id": "window.4", "bbox": [530, 243, 720, 391]}, {"id": "train.5", "bbox": [12, 187, 1019, 547]}, {"id": "letter.6", "bbox": [430, 339, 469, 398]}, {"id": "train.7", "bbox": [499, 205, 1010, 538]}, {"id": "window.8", "bbox": [479, 286, 559, 373]}, {"id": "window.9", "bbox": [57, 257, 224, 393]}, {"id": "window.10", "bbox": [996, 315, 1021, 401]}, {"id": "window.11", "bbox": [783, 296, 811, 415]}, {"id": "window.12", "bbox": [336, 290, 371, 344]}, {"id": "train.13", "bbox": [15, 202, 553, 610]}, {"id": "train.14", "bbox": [14, 212, 1015, 580]}, {"id": "window.15", "bbox": [512, 233, 725, 469]}]
|
[{"subject": "window.9", "predicate": "on", "object": "train.13"}, {"subject": "window.12", "predicate": "on", "object": "train.13"}, {"subject": "window.4", "predicate": "on", "object": "train.7"}, {"subject": "window.11", "predicate": "on", "object": "train.7"}, {"subject": "window.10", "predicate": "on", "object": "train.7"}, {"subject": "window.8", "predicate": "on", "object": "train.13"}, {"subject": "window.8", "predicate": "on", "object": "train.3"}, {"subject": "letter.6", "predicate": "on", "object": "train.3"}, {"subject": "window.15", "predicate": "on", "object": "train.7"}]
|
|
498364
|
Generate a structured scene graph for an image of size (1021 x 1024) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1021 x 1024) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "pole.1", "bbox": [406, 0, 459, 707]}, {"id": "sign.2", "bbox": [576, 504, 782, 703]}, {"id": "sign.3", "bbox": [509, 563, 566, 621]}, {"id": "sign.4", "bbox": [117, 495, 312, 675]}, {"id": "truck.5", "bbox": [729, 485, 867, 632]}, {"id": "car.6", "bbox": [793, 511, 1015, 687]}]
|
[{"subject": "car.6", "predicate": "behind", "object": "truck.5"}]
|
|
498365
|
Generate a structured scene graph for an image of size (1000 x 750) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1000 x 750) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "bike.1", "bbox": [364, 380, 522, 528]}, {"id": "bike.2", "bbox": [715, 479, 998, 745]}, {"id": "boy.3", "bbox": [373, 316, 458, 521]}, {"id": "leaf.4", "bbox": [294, 133, 333, 204]}, {"id": "leaf.5", "bbox": [726, 295, 758, 381]}, {"id": "leaf.6", "bbox": [500, 283, 583, 364]}, {"id": "leaf.7", "bbox": [675, 299, 724, 366]}, {"id": "leaf.8", "bbox": [609, 294, 641, 364]}, {"id": "leaf.9", "bbox": [347, 106, 410, 172]}, {"id": "shirt.10", "bbox": [385, 342, 451, 420]}, {"id": "short.11", "bbox": [397, 414, 447, 471]}, {"id": "sign.12", "bbox": [458, 351, 683, 414]}, {"id": "tire.13", "bbox": [715, 570, 894, 745]}, {"id": "tire.14", "bbox": [449, 445, 521, 526]}, {"id": "tree.15", "bbox": [672, 295, 758, 463]}, {"id": "tree.16", "bbox": [456, 115, 641, 262]}, {"id": "tree.17", "bbox": [486, 267, 581, 355]}]
|
[{"subject": "boy.3", "predicate": "wearing", "object": "shirt.10"}, {"subject": "leaf.6", "predicate": "on", "object": "tree.17"}, {"subject": "leaf.5", "predicate": "on", "object": "tree.15"}]
|
|
498366
|
Generate a structured scene graph for an image of size (960 x 1280) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (960 x 1280) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "clock.1", "bbox": [323, 490, 468, 640]}, {"id": "leg.2", "bbox": [236, 953, 295, 1277]}, {"id": "plant.3", "bbox": [310, 31, 635, 361]}, {"id": "pot.4", "bbox": [428, 217, 547, 292]}, {"id": "shelf.5", "bbox": [213, 281, 783, 1278]}, {"id": "toilet.6", "bbox": [316, 1017, 681, 1273]}, {"id": "towel.7", "bbox": [420, 856, 687, 982]}, {"id": "towel.8", "bbox": [405, 941, 686, 1073]}, {"id": "towel.9", "bbox": [400, 730, 690, 911]}, {"id": "vase.10", "bbox": [433, 513, 525, 695]}, {"id": "vase.11", "bbox": [583, 498, 645, 672]}, {"id": "toilet.12", "bbox": [328, 1045, 617, 1273]}, {"id": "handle.13", "bbox": [325, 1085, 375, 1142]}]
|
[{"subject": "shelf.5", "predicate": "above", "object": "toilet.6"}, {"subject": "plant.3", "predicate": "above", "object": "shelf.5"}, {"subject": "pot.4", "predicate": "on", "object": "shelf.5"}, {"subject": "clock.1", "predicate": "on", "object": "shelf.5"}, {"subject": "towel.9", "predicate": "on", "object": "shelf.5"}, {"subject": "towel.7", "predicate": "on", "object": "shelf.5"}, {"subject": "towel.8", "predicate": "on", "object": "shelf.5"}, {"subject": "vase.10", "predicate": "on", "object": "shelf.5"}, {"subject": "vase.11", "predicate": "on", "object": "shelf.5"}, {"subject": "handle.13", "predicate": "on", "object": "toilet.12"}]
|
|
498367
|
Generate a structured scene graph for an image of size (675 x 1024) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (675 x 1024) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "bag.1", "bbox": [318, 527, 439, 768]}, {"id": "bag.2", "bbox": [85, 545, 150, 756]}, {"id": "building.3", "bbox": [0, 4, 675, 639]}, {"id": "hair.4", "bbox": [339, 453, 403, 518]}, {"id": "jacket.5", "bbox": [0, 534, 115, 725]}, {"id": "jacket.6", "bbox": [438, 505, 570, 659]}, {"id": "jacket.7", "bbox": [545, 516, 673, 670]}, {"id": "jacket.8", "bbox": [314, 509, 468, 685]}, {"id": "man.9", "bbox": [70, 481, 135, 808]}, {"id": "man.10", "bbox": [437, 455, 579, 828]}, {"id": "pant.11", "bbox": [0, 706, 105, 908]}, {"id": "pant.12", "bbox": [353, 749, 434, 858]}, {"id": "shirt.13", "bbox": [135, 549, 208, 614]}, {"id": "shirt.14", "bbox": [72, 514, 132, 610]}, {"id": "shoe.15", "bbox": [41, 901, 80, 940]}, {"id": "tie.16", "bbox": [477, 508, 500, 606]}, {"id": "woman.17", "bbox": [135, 506, 226, 686]}, {"id": "woman.18", "bbox": [309, 454, 540, 946]}, {"id": "woman.19", "bbox": [522, 454, 674, 884]}, {"id": "woman.20", "bbox": [0, 470, 152, 952]}, {"id": "tie.21", "bbox": [471, 505, 522, 605]}, {"id": "shoe.22", "bbox": [337, 870, 420, 941]}]
|
[{"subject": "man.10", "predicate": "wearing", "object": "tie.21"}, {"subject": "woman.20", "predicate": "wearing", "object": "pant.11"}, {"subject": "woman.18", "predicate": "wearing", "object": "jacket.8"}, {"subject": "woman.18", "predicate": "wearing", "object": "pant.12"}, {"subject": "woman.18", "predicate": "holding", "object": "bag.1"}, {"subject": "man.10", "predicate": "wearing", "object": "tie.16"}, {"subject": "woman.18", "predicate": "wearing", "object": "shoe.22"}, {"subject": "woman.20", "predicate": "holding", "object": "bag.2"}, {"subject": "woman.20", "predicate": "wearing", "object": "jacket.5"}, {"subject": "woman.20", "predicate": "wearing", "object": "shoe.15"}, {"subject": "woman.18", "predicate": "with", "object": "hair.4"}, {"subject": "man.10", "predicate": "in", "object": "jacket.6"}, {"subject": "man.9", "predicate": "in", "object": "shirt.14"}, {"subject": "woman.17", "predicate": "in", "object": "shirt.13"}, {"subject": "jacket.8", "predicate": "on", "object": "woman.18"}, {"subject": "jacket.7", "predicate": "on", "object": "woman.19"}]
|
|
498368
|
Generate a structured scene graph for an image of size (1024 x 363) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 363) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "curtain.1", "bbox": [211, 69, 324, 232]}, {"id": "lamp.2", "bbox": [753, 110, 804, 185]}, {"id": "room.3", "bbox": [0, 0, 1024, 363]}, {"id": "glass.4", "bbox": [397, 249, 464, 332]}, {"id": "table.5", "bbox": [317, 269, 756, 352]}]
|
[{"subject": "glass.4", "predicate": "on", "object": "table.5"}]
|
|
498369
|
Generate a structured scene graph for an image of size (1024 x 683) using the following format:
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign a unique ID for each object using the format `"object_name.number"` (e.g., `"person.1"`, `"bike.2"`).
- Provide its bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Represent interactions accurately using `"subject"`, `"predicate"`, and `"object"`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
Generate a structured scene graph for an image of size (1024 x 683) using the specified object and relationship categories.
### **Output Format:**
<answer>
{
"objects": [
{"id": "object_name.number", "bbox": [x1, y1, x2, y2]},
...
],
"relationships": [
{"subject": "object_name.number", "predicate": "relationship_type", "object": "object_name.number"},
...
]
}
</answer>
### **Guidelines:**
- **Objects:**
- Assign unique IDs in the format `"object_name.number"` (e.g., `"person.1"`). The **object_name** must belong to the predefined object set: `["airplane", "animal", "arm", "bag", "banana", "basket", "beach", "bear", "bed", "bench", "bike", "bird", "board", "boat", "book", "boot", "bottle", "bowl", "box", "boy", "branch", "building", "bus", "cabinet", "cap", "car", "cat", "chair", "child", "clock", "coat", "counter", "cow", "cup", "curtain", "desk", "dog", "door", "drawer", "ear", "elephant", "engine", "eye", "face", "fence", "finger", "flag", "flower", "food", "fork", "fruit", "giraffe", "girl", "glass", "glove", "guy", "hair", "hand", "handle", "hat", "head", "helmet", "hill", "horse", "house", "jacket", "jean", "kid", "kite", "lady", "lamp", "laptop", "leaf", "leg", "letter", "light", "logo", "man", "men", "motorcycle", "mountain", "mouth", "neck", "nose", "number", "orange", "pant", "paper", "paw", "people", "person", "phone", "pillow", "pizza", "plane", "plant", "plate", "player", "pole", "post", "pot", "racket", "railing", "rock", "roof", "room", "screen", "seat", "sheep", "shelf", "shirt", "shoe", "short", "sidewalk", "sign", "sink", "skateboard", "ski", "skier", "sneaker", "snow", "sock", "stand", "street", "surfboard", "table", "tail", "tie", "tile", "tire", "toilet", "towel", "tower", "track", "train", "tree", "truck", "trunk", "umbrella", "vase", "vegetable", "vehicle", "wave", "wheel", "window", "windshield", "wing", "wire", "woman", "zebra"]`.
- Provide a bounding box `[x1, y1, x2, y2]` in integer pixel format.
- Include all visible objects, even if they have no relationships.
- **Relationships:**
- Define relationships using `"subject"`, `"predicate"`, and `"object"`.
- The **predicate** must belong to the predefined relationship set: `["above", "across", "against", "along", "and", "at", "attached to", "behind", "belonging to", "between", "carrying", "covered in", "covering", "eating", "flying in", "for", "from", "growing on", "hanging from", "has", "holding", "in", "in front of", "laying on", "looking at", "lying on", "made of", "mounted on", "near", "of", "on", "on back of", "over", "painted on", "parked on", "part of", "playing", "riding", "says", "sitting on", "standing on", "to", "under", "using", "walking in", "walking on", "watching", "wearing", "wears", "with"]`.
- Omit relationships for orphan objects.
### **Example Output:**
<answer>
{
"objects": [
{"id": "person.1", "bbox": [120, 200, 350, 700]},
{"id": "bike.2", "bbox": [100, 600, 400, 800]},
{"id": "helmet.3", "bbox": [150, 150, 280, 240]},
{"id": "tree.4", "bbox": [500, 100, 750, 700]}
],
"relationships": [
{"subject": "person.1", "predicate": "riding", "object": "bike.2"},
{"subject": "person.1", "predicate": "wearing", "object": "helmet.3"}
]
}
</answer>
Now, generate the complete scene graph for the provided image:
|
[{"id": "arm.1", "bbox": [252, 190, 326, 234]}, {"id": "bag.2", "bbox": [347, 243, 477, 351]}, {"id": "boy.3", "bbox": [342, 382, 749, 540]}, {"id": "chair.4", "bbox": [570, 423, 846, 657]}, {"id": "face.5", "bbox": [579, 73, 625, 127]}, {"id": "lady.6", "bbox": [163, 155, 378, 468]}, {"id": "pant.7", "bbox": [256, 350, 405, 484]}, {"id": "person.8", "bbox": [517, 51, 688, 202]}, {"id": "table.9", "bbox": [318, 139, 1005, 524]}, {"id": "tile.10", "bbox": [101, 489, 397, 652]}, {"id": "boy.11", "bbox": [327, 403, 481, 529]}, {"id": "shirt.12", "bbox": [402, 423, 575, 517]}, {"id": "boy.13", "bbox": [331, 380, 593, 540]}, {"id": "cup.14", "bbox": [654, 268, 697, 333]}, {"id": "plate.15", "bbox": [758, 305, 819, 360]}, {"id": "plate.16", "bbox": [799, 280, 860, 324]}, {"id": "plate.17", "bbox": [803, 302, 866, 374]}, {"id": "woman.18", "bbox": [485, 53, 684, 194]}, {"id": "woman.19", "bbox": [626, 43, 841, 227]}, {"id": "table.20", "bbox": [372, 157, 943, 559]}, {"id": "seat.21", "bbox": [657, 440, 770, 546]}, {"id": "shirt.22", "bbox": [218, 211, 353, 397]}, {"id": "guy.23", "bbox": [324, 372, 485, 539]}, {"id": "woman.24", "bbox": [196, 148, 329, 299]}, {"id": "child.25", "bbox": [157, 199, 285, 374]}, {"id": "child.26", "bbox": [164, 189, 247, 304]}, {"id": "shirt.27", "bbox": [430, 437, 606, 520]}, {"id": "chair.28", "bbox": [409, 489, 656, 656]}, {"id": "bag.29", "bbox": [331, 231, 472, 320]}, {"id": "shirt.30", "bbox": [530, 98, 664, 202]}, {"id": "plate.31", "bbox": [495, 212, 562, 271]}, {"id": "food.32", "bbox": [499, 235, 554, 268]}]
|
[{"subject": "boy.11", "predicate": "wearing", "object": "shirt.12"}, {"subject": "cup.14", "predicate": "sitting on", "object": "table.9"}, {"subject": "chair.4", "predicate": "with", "object": "seat.21"}, {"subject": "lady.6", "predicate": "wearing", "object": "shirt.22"}, {"subject": "guy.23", "predicate": "wearing", "object": "shirt.12"}, {"subject": "table.9", "predicate": "near", "object": "woman.24"}, {"subject": "woman.24", "predicate": "holding", "object": "child.25"}, {"subject": "woman.24", "predicate": "holding", "object": "child.26"}, {"subject": "woman.24", "predicate": "with", "object": "arm.1"}, {"subject": "boy.3", "predicate": "wearing", "object": "shirt.27"}, {"subject": "bag.29", "predicate": "on", "object": "table.9"}, {"subject": "person.8", "predicate": "in", "object": "shirt.30"}, {"subject": "plate.31", "predicate": "of", "object": "food.32"}]
|
End of preview. Expand
in Data Studio
This repository contains the VG150 dataset transformed into datasets format with keys: "image_id", "image", "prompt_open", "prompt_close", "objects", and "relationships".
It was presented in the paper Compile Scene Graphs with Reinforcement Learning.
- Downloads last month
- 362
Size of downloaded dataset files:
8.14 GB
Size of the auto-converted Parquet files:
8.14 GB
Number of rows:
56,224