Datasets:

OpenIXCLab
/

SeCVOS

video_id stringlengths 11 11	obj_id stringclasses 4 values	num_frames int64 48 339	exp_text stringlengths 11 117
-klPo9pilc4	000	135	A rectangular cardboard box.
-klPo9pilc4	001	135	A brown cardboard tube.
-klPo9pilc4	002	135	A yellow utility knife.
-nIt1BX3DFY	000	185	The red Ford F-150 Raptor pickup truck.
09trCkd99Os	000	128	A teal Volkswagen T-Cross SUV.
0D27yQy-qWk	000	149	A red and black bicycle.
0HgnueqXKyU	000	185	An orange Dodge Viper sports car.
0Jn5N_pmZuA	000	158	A dark-colored sedan.
0SVMNRIKTaw	000	137	A motorcycle rider in a helmet and motorcycle gear.
0SVMNRIKTaw	001	137	A yellow and black sport motorcycle.
0ZcDkQX6WFE	000	204	A superhero in a red and silver suit with a dinosaur-themed helmet.
0kibzqZ4Fe8	000	127	A woman in a colorful unicorn onesie.
0waHmuWSPZo	000	88	A man in a black hooded sweatshirt and headphones.
1AtGI9kSzpY	000	126	A young woman with long blonde hair.
290cMhsTCVM	000	160	A white Subaru WRX sedan.
2UNKqQHG4gA	000	154	A brown-haired animated bride in a white dress.
2UNKqQHG4gA	001	154	An elderly man with spiky blue-gray hair wearing a white lab coat.
2yFGHfV4L6U	000	157	A female tennis player in a red shirt and purple headband.
3qIRJLGGZy0	000	339	A grey cat in a police uniform and hat.
3qIRJLGGZy0	001	339	A tabby cat in a blue shirt and large gold chain.
42GofA15N3M	000	128	A long-haired blonde man performing with an acoustic guitar.
4AxWuk52ZUQ	000	217	A man in sunglasses.
4glQyjzXgrE	000	120	A Black man in a backward cap and a light-colored zip-up jacket.
4glQyjzXgrE	001	120	A blond-haired man in a light-colored t-shirt featuring a dolphin graphic.
4i-xiG8a13Q	000	154	A crocodile.
4xB9ab4QUiw	000	195	A cup of yogurt.
5a45aSpta88	000	150	A bearded man in a grey blazer.
5a45aSpta88	001	150	A six-wheeled white and black autonomous delivery robot.
63HFjBi0oAQ	000	166	A dark-colored McLaren sports car.
650oOOyEKG4	000	251	A light blue-collared animated brown puppy.
6BIvQHzdIC0	000	120	A black dog.
6BIvQHzdIC0	001	120	A elephant.
6ihbDGFbajc	000	121	An emergency pack with "i survived" printed on it..
6mOZBTdzOj0	000	170	A man in a black jacket and white backward baseball cap.
854xZGo5pQw	000	152	A young man with short dark hair in a red button-up shirt and blue jeans.
8CUa81t4H5s	000	158	A sailboat with white sails.
8a6_6LWj3FY	000	217	A video game character in cowboy outfit.
9-P4pMS6F2c	000	195	A dark grey Aston Martin sports car.
AeP2Dwree4M	000	251	The drum in front of the boy.
BPNi5dTHwKM	000	100	A long-haired female bike polo player in a purple tank top and a backward black baseball cap.
BU-wIj--VWs	000	128	A blue BMW M2 coupe.
BpSmYQvR8cA	000	183	A medium-sized wild cat with a spotted coat.
BuRf6r0LuL8	000	323	A lime green, black, and white striped wakesurf board.
C-NDP5g2ERQ	000	155	A young woman with blonde hair in a dark suit jacket, white collared shirt, and tie.
C5ELpD4OZGM	000	177	A young man with dark hair in a two-toned jacket over a plaid shirt.
CKmkgzeNKZw	000	191	A pair of black headphones.
CKrV4mwTka0	000	200	A man wearing a black tank top and shorts.
Cgj-gtEj8fU	000	173	A gold-colored round doorknob.
Crf2aApb0yw	000	228	A white Fiat 500 Abarth with a red side stripe.
DBDUDgCG-94	000	173	A bearded man in a grey jacket and scarf.
DBDUDgCG-94	001	173	An open-door red sports car.
DGcdKA7159E	000	146	A man in a red bandana.
DebmDk-weKs	000	133	A man in a red shirt.
DqC3ieJJlFM	000	179	A pangolin.
EFgdx_t4baQ	000	164	A white Peugeot 508 hybrid car.
EO-KsZYeHRA	000	150	A grey animated dog in a green helmet and a blue and green wetsuit.
E_UbFKdWFl4	000	161	A black Samsung Galaxy Note10 smartphone.
EoQaVBlutWw	000	147	The racing white car.
FIIKq3MBCg0	000	151	A woman with short dark hair in a blue shirt and pink apron.
FIIKq3MBCg0	001	151	A silver metal mixing bowl.
FRNZQHmo_Ig	000	251	A black and grey animated dog in a green and white polka-dotted party hat.
FRNZQHmo_Ig	001	251	A cartoon light brown dog with a tan muzzle wearing a red and white striped party hat.
FRNZQHmo_Ig	002	251	A yellow cartoon dog in black sunglasses and a red and white striped party hat.
FRNZQHmo_Ig	003	251	A dancing cartoon grey dog in a green and white striped party hat.
FhkXz51WsCg	000	188	A light blue vintage Shelby GT500 KR convertible car.
FmHI0ar0EvE	000	122	A red Maruti Vitara Brezza car.
FsLaTZmP6Uw	000	265	The table tennis racket in the hand of the player wearing a red outfit.
FsLaTZmP6Uw	001	265	The table tennis racket in the hand of the player wearing a black outfit.
GDoI16RmxqQ	000	128	A silver Rivian electric pickup truck.
GKqdC_xcf28	000	152	A woman with reddish-brown hair, a patterned top.
GMnti60Gxgc	000	195	A brown and beige plaid fabric.
Gcmn73sImtQ	000	168	A young girl in a red and black polka-dotted Ladybug costume.
Gq-DY2iiqbc	000	130	A brown and white bulldog.
H7hkYP9CeSk	000	179	A green and white #18 NASCAR stock car.
HC-XSzO3rrE	000	151	A yellow combine harvester.
HPTVnL0FN2o	000	225	A small, white rolling pin.
HmHjI9Z_vbI	000	77	An Emirates Airbus A380 airplane.
IR6V_6xqWDU	000	220	A small, portable external speaker with a transparent casing and a silver cone.
IaiKA_fh_q0	000	180	A white wall-pluggable Zyxel powerline network adapter.
IppaafjwUUA	000	155	A brown and yellow short-haired dog.
JAkIt2GhWF0	000	135	A small, orange Tata Nano compact car.
JF7co5M06kY	000	173	A cartoon pig character dressed in a pirate costume with a black hat, a red shirt, and a gold earring.
JJNUDzMqeFU	000	195	a silver SUV.
JTtcExfeKi0	000	150	An older man with white hair and glasses in a green button-up shirt and blue jeans.
Jo9zeWfXYHA	000	241	A red saucepan of liquid.
K9U5YEQy1H8	000	123	The portrayal of a groundskeeper wearing a camouflage bucket hat.
KHdxZ-p1Uho	000	192	An animated red monster truck equipped with a large articulated arm holding a saw.
KKxyz7YPED4	000	95	A homemade black long-sleeved Christmas sweater with a picture of Shrek in a Santa hat and the pun "Shrek the Halls."
KNAXihfJirw	000	197	A young girl with tied-back blonde hair.
KYt8h7ydprA	000	184	A female figure skater.
K_cVTBY0NoE	000	188	A red Ferrari 512 TR.
Ki4QY8G6eWo	000	255	A black glue stick with a white cap.
LFpk448qjKQ	000	145	A baby elephant.
M0WeiG2-HRQ	000	119	A female anime character with short black hair and a red hooded cape.
MDDtQoaj6sA	000	166	A blue Toyota Tacoma.
MDDtQoaj6sA	001	166	A red pickup truck.
MF7FK-3Kg3A	000	131	A TV remote control.
MTQGzgPMqoc	000	135	The mounted gray, weatherproof electrical junction box.
N_KKXG5NZdg	000	168	A blue My Little Pony toy with a rainbow-colored mane.
NdV-EAKV0Hk	000	48	The yellow animated bird from the Angry Birds franchise.

End of preview. Expand in Data Studio

SeC: Advancing Complex Video Object Segmentation via Progressive Concept Construction

[📂 GitHub] [📦 Model] [🌐 Homepage] [📄 Paper]

Highlights

🔥We introduce Segment Concept (SeC), a concept-driven segmentation framework for video object segmentation that integrates Large Vision-Language Models (LVLMs) for robust, object-centric representations.
🔥SeC dynamically balances semantic reasoning with feature matching, adaptively adjusting computational efforts based on scene complexity for optimal segmentation performance.
🔥We propose the Semantic Complex Scenarios Video Object Segmentation (SeCVOS) benchmark, designed to evaluate segmentation in challenging scenarios.

SeCVOS Benchmark

We propose the Semantic Complex Scenarios Video Object Segmentation (SeCVOS) benchmark, specifically designed to assess a model’s ability to perform high-level semantic reasoning across complex visual narratives. SeCVOS contains 160 carefully curated multi-shot videos characterized by: 1) Highly discontinuous frame sequences, 2) Frequent reappearance of objects across disparate scenes, and 3) Abrupt shot transitions and dynamic camera motion.

Benchmark	#Videos	Avg. Duration (s)	Disapp. Rate	Avg. #Scene
DAVIS	90	2.87	16.1%	1.06
YTVOS	507	4.51	13.0%	1.03
MOSE	311	8.68*	41.5%	1.06
SA-V	155	17.24	25.5%	1.09
LVOS	140	78.36	7.8%	1.47
SeCVOS (ours)	160	29.36	30.2%	4.26

License

Our annotations are licensed under a CC BY-NC-SA 4.0 License. They are available strictly for non-commercial research.

We uphold the rights of individuals and copyright holders. If you are featured in any of our video annotations or hold copyright to a video and wish to have its annotation removed from our dataset, please reach out to us. Send an email to [email protected] with the subject line beginning with SeCVOS, or raise an issue with the same title format. We commit to reviewing your request promptly and taking suitable action.

Citation

If you find this project useful in your research, please consider citing:

@article{zhang2025sec,
  title     = {SeC: Advancing Complex Video Object Segmentation via Progressive Concept Construction},
  author    = {Zhixiong Zhang and Shuangrui Ding and Xiaoyi Dong and Songxin He and Jianfan Lin and Junsong Tang and Yuhang Zang and Yuhang Cao and Dahua Lin and Jiaqi Wang},
  journal   = {arXiv preprint arXiv:2507.15852},
  year      = {2025}
}

Downloads last month: 51

Size of downloaded dataset files:

22.6 kB

Size of the auto-converted Parquet files:

10.8 kB

Number of rows:

182