Update README.md
Browse files
README.md
CHANGED
|
@@ -8,4 +8,62 @@ metrics:
|
|
| 8 |
base_model:
|
| 9 |
- Ultralytics/YOLOv8
|
| 10 |
pipeline_tag: object-detection
|
| 11 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
base_model:
|
| 9 |
- Ultralytics/YOLOv8
|
| 10 |
pipeline_tag: object-detection
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# YOLO Document Layout Model
|
| 14 |
+
|
| 15 |
+
This model is a fine-tuned YOLO detector for document layout analysis, capable of identifying various document elements such as text columns, figures, tables, and other typographical features.
|
| 16 |
+
|
| 17 |
+
## Model Description
|
| 18 |
+
|
| 19 |
+
The model is trained to detect and classify 20 different document components, including text structures (TextColumn, List), semantic elements (Title, Header), typographical features (Bold, Italic), and visual components (Figure, Table).
|
| 20 |
+
|
| 21 |
+
## Model Detections
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
### Training
|
| 28 |
+
|
| 29 |
+
The model was fine-tuned using a proprietary dataset of document images.
|
| 30 |
+
|
| 31 |
+
## Evaluation Results
|
| 32 |
+
|
| 33 |
+
The model's performance was evaluated on a test set with the following metrics:
|
| 34 |
+
|
| 35 |
+
| Class | Images | Instances | Precision | Recall | mAP50 | mAP50-95 |
|
| 36 |
+
|-------|--------|-----------|-----------|--------|-------|----------|
|
| 37 |
+
| **all** | **150** | **1255** | **0.701** | **0.723** | **0.735** | **0.509** |
|
| 38 |
+
| Author | 7 | 65 | 0.693 | 0.174 | 0.307 | 0.134 |
|
| 39 |
+
| Bigletter | 11 | 11 | 1.000 | 0.900 | 0.976 | 0.563 |
|
| 40 |
+
| Bleeding | 9 | 10 | 0.618 | 0.700 | 0.667 | 0.547 |
|
| 41 |
+
| Bold | 23 | 77 | 0.679 | 0.753 | 0.798 | 0.395 |
|
| 42 |
+
| Caption | 50 | 71 | 0.892 | 0.816 | 0.881 | 0.642 |
|
| 43 |
+
| Date | 17 | 57 | 0.927 | 0.666 | 0.728 | 0.386 |
|
| 44 |
+
| Figure | 90 | 149 | 0.772 | 0.725 | 0.823 | 0.677 |
|
| 45 |
+
| Footnote | 14 | 15 | 0.500 | 0.667 | 0.612 | 0.478 |
|
| 46 |
+
| Header | 16 | 16 | 0.560 | 0.717 | 0.664 | 0.476 |
|
| 47 |
+
| Italic | 17 | 86 | 0.448 | 0.791 | 0.557 | 0.327 |
|
| 48 |
+
| List | 34 | 55 | 0.615 | 0.709 | 0.742 | 0.591 |
|
| 49 |
+
| Map | 4 | 4 | 0.606 | 0.750 | 0.656 | 0.599 |
|
| 50 |
+
| SubSubTitle | 37 | 97 | 0.627 | 0.520 | 0.599 | 0.300 |
|
| 51 |
+
| SubTitle | 54 | 96 | 0.605 | 0.562 | 0.605 | 0.327 |
|
| 52 |
+
| Table | 30 | 43 | 0.865 | 0.953 | 0.966 | 0.855 |
|
| 53 |
+
| TextColumn | 115 | 323 | 0.831 | 0.913 | 0.933 | 0.811 |
|
| 54 |
+
| Title | 47 | 66 | 0.712 | 0.711 | 0.649 | 0.441 |
|
| 55 |
+
| Underline | 2 | 4 | 0.681 | 1.000 | 0.995 | 0.665 |
|
| 56 |
+
| equations | 4 | 10 | 0.688 | 0.700 | 0.809 | 0.450 |
|
| 57 |
+
|
| 58 |
+
### Key Performance Highlights:
|
| 59 |
+
|
| 60 |
+
- **Best performing classes**: Table (mAP50: 0.966), TextColumn (mAP50: 0.933), and Caption (mAP50: 0.881)
|
| 61 |
+
- **High precision classes**: Bigletter (1.000), Date (0.927), and Caption (0.892)
|
| 62 |
+
- **High recall classes**: Underline (1.000), Table (0.953), and TextColumn (0.913)
|
| 63 |
+
- **Overall performance**: mAP50 of 0.735 and mAP50-95 of 0.509 across all classes
|
| 64 |
+
|
| 65 |
+
## Limitations
|
| 66 |
+
|
| 67 |
+
- Lower performance on Author detection (mAP50: 0.307)
|
| 68 |
+
- Moderate performance on typographical features like Italic (mAP50: 0.557)
|
| 69 |
+
- Limited sample size for some classes (Map, Underline, equations)
|