Datasets:
update readme with nes visualizations
Browse files- README.md +21 -14
- analysis/reading_order_sample_1.png +2 -2
- analysis/reading_order_sample_10.png +3 -0
- analysis/reading_order_sample_2.png +2 -2
- analysis/reading_order_sample_3.png +2 -2
- analysis/reading_order_sample_4.png +2 -2
- analysis/reading_order_sample_5.png +3 -0
- analysis/reading_order_sample_6.png +3 -0
- analysis/reading_order_sample_7.png +3 -0
- analysis/reading_order_sample_8.png +3 -0
- analysis/reading_order_sample_9.png +3 -0
- analysis/reading_order_visualization.png +2 -2
README.md
CHANGED
|
@@ -26,18 +26,6 @@ This dataset benchmarks reading order detection models on complex, real-world do
|
|
| 26 |
<img src="analysis/dataset_overview.png" alt="Dataset Overview" width="100%"/>
|
| 27 |
</div>
|
| 28 |
|
| 29 |
-
### Reading Order Annotations
|
| 30 |
-
|
| 31 |
-
<div align="center">
|
| 32 |
-
<img src="analysis/reading_order_visualization.png" alt="Reading Order Visualization" width="100%"/>
|
| 33 |
-
</div>
|
| 34 |
-
|
| 35 |
-
The dataset provides **professionally annotated reading order sequences** for each document, capturing the natural flow in which humans would read the content. This is critical for:
|
| 36 |
-
- **Accurate text extraction** from complex layouts
|
| 37 |
-
- **Document understanding** and information retrieval
|
| 38 |
-
- **Maintaining semantic coherence** in multi-column documents
|
| 39 |
-
- **Preserving logical flow** across mixed layout elements
|
| 40 |
-
|
| 41 |
### Document Categories
|
| 42 |
|
| 43 |
- Financial (14.3%), Legal (13.1%), Medical (10.8%), Government (10.0%)
|
|
@@ -108,9 +96,13 @@ chunkr-reading-order-bench-oss/
|
|
| 108 |
|
| 109 |
---
|
| 110 |
|
| 111 |
-
|
| 112 |
|
| 113 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 114 |
|
| 115 |
<div align="center">
|
| 116 |
<img src="analysis/reading_order_sample_1.png" alt="Single Column Example" width="45%"/>
|
|
@@ -122,6 +114,21 @@ chunkr-reading-order-bench-oss/
|
|
| 122 |
<img src="analysis/reading_order_sample_4.png" alt="Mixed Elements Example" width="45%"/>
|
| 123 |
</div>
|
| 124 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 125 |
The arrows indicate the ground truth reading order progression, showing how the document should be read to maintain semantic coherence.
|
| 126 |
|
| 127 |
---
|
|
|
|
| 26 |
<img src="analysis/dataset_overview.png" alt="Dataset Overview" width="100%"/>
|
| 27 |
</div>
|
| 28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 29 |
### Document Categories
|
| 30 |
|
| 31 |
- Financial (14.3%), Legal (13.1%), Medical (10.8%), Government (10.0%)
|
|
|
|
| 96 |
|
| 97 |
---
|
| 98 |
|
| 99 |
+
### Reading Order Annotations
|
| 100 |
|
| 101 |
+
The dataset provides **professionally annotated reading order sequences** for each document, capturing the natural flow in which humans would read the content. This is critical for:
|
| 102 |
+
- **Accurate text extraction** from complex layouts
|
| 103 |
+
- **Document understanding** and information retrieval
|
| 104 |
+
- **Maintaining semantic coherence** in multi-column documents
|
| 105 |
+
- **Preserving logical flow** across mixed layout elements
|
| 106 |
|
| 107 |
<div align="center">
|
| 108 |
<img src="analysis/reading_order_sample_1.png" alt="Single Column Example" width="45%"/>
|
|
|
|
| 114 |
<img src="analysis/reading_order_sample_4.png" alt="Mixed Elements Example" width="45%"/>
|
| 115 |
</div>
|
| 116 |
|
| 117 |
+
<div align="center">
|
| 118 |
+
<img src="analysis/reading_order_sample_5.png" alt="Complex Layout Example" width="45%"/>
|
| 119 |
+
<img src="analysis/reading_order_sample_6.png" alt="Mixed Elements Example" width="45%"/>
|
| 120 |
+
</div>
|
| 121 |
+
|
| 122 |
+
<div align="center">
|
| 123 |
+
<img src="analysis/reading_order_sample_7.png" alt="Complex Layout Example" width="45%"/>
|
| 124 |
+
<img src="analysis/reading_order_sample_8.png" alt="Mixed Elements Example" width="45%"/>
|
| 125 |
+
</div>
|
| 126 |
+
|
| 127 |
+
<div align="center">
|
| 128 |
+
<img src="analysis/reading_order_sample_9.png" alt="Complex Layout Example" width="45%"/>
|
| 129 |
+
<img src="analysis/reading_order_sample_10.png" alt="Mixed Elements Example" width="45%"/>
|
| 130 |
+
</div>
|
| 131 |
+
|
| 132 |
The arrows indicate the ground truth reading order progression, showing how the document should be read to maintain semantic coherence.
|
| 133 |
|
| 134 |
---
|
analysis/reading_order_sample_1.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|
analysis/reading_order_sample_10.png
ADDED
|
Git LFS Details
|
analysis/reading_order_sample_2.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|
analysis/reading_order_sample_3.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|
analysis/reading_order_sample_4.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|
analysis/reading_order_sample_5.png
ADDED
|
Git LFS Details
|
analysis/reading_order_sample_6.png
ADDED
|
Git LFS Details
|
analysis/reading_order_sample_7.png
ADDED
|
Git LFS Details
|
analysis/reading_order_sample_8.png
ADDED
|
Git LFS Details
|
analysis/reading_order_sample_9.png
ADDED
|
Git LFS Details
|
analysis/reading_order_visualization.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|