Datasets:
				
			
			
	
			
	
		
			
	
		
		Update README.md
Browse files
    	
        README.md
    CHANGED
    
    | 
         @@ -1,3 +1,57 @@ 
     | 
|
| 1 | 
         
            -
            ---
         
     | 
| 2 | 
         
            -
            license: gpl-3.0
         
     | 
| 3 | 
         
            -
             
     | 
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
|
| 
         | 
| 
         | 
|
| 1 | 
         
            +
            ---
         
     | 
| 2 | 
         
            +
            license: gpl-3.0
         
     | 
| 3 | 
         
            +
            tags:
         
     | 
| 4 | 
         
            +
              - human-pose-estimation
         
     | 
| 5 | 
         
            +
              - computer-vision
         
     | 
| 6 | 
         
            +
              - person-detection
         
     | 
| 7 | 
         
            +
              - amodal-person-detection
         
     | 
| 8 | 
         
            +
            size_categories:
         
     | 
| 9 | 
         
            +
              - 1K<n<10K
         
     | 
| 10 | 
         
            +
            task_categories:
         
     | 
| 11 | 
         
            +
              - keypoint-detection
         
     | 
| 12 | 
         
            +
              - object-detection
         
     | 
| 13 | 
         
            +
            language:
         
     | 
| 14 | 
         
            +
              - en
         
     | 
| 15 | 
         
            +
            ---
         
     | 
| 16 | 
         
            +
             
     | 
| 17 | 
         
            +
            # CropCOCO Dataset
         
     | 
| 18 | 
         
            +
             
     | 
| 19 | 
         
            +
            CropCOCO is a validation-only dataset of COCO val 2017 images cropped such that some keypoints annotations are outside of the image.
         
     | 
| 20 | 
         
            +
            It can be used for keypoint detection, out-of-image keypoint detection and localization, person detection and amodal person detection.
         
     | 
| 21 | 
         
            +
             
     | 
| 22 | 
         
            +
            [](https://github.com/MiraPurkrabek/ProbPose_code)
         
     | 
| 23 | 
         
            +
            [](https://arxiv.org/abs/2412.02254)
         
     | 
| 24 | 
         
            +
            [](https://mirapurkrabek.github.io/ProbPose/)
         
     | 
| 25 | 
         
            +
             
     | 
| 26 | 
         
            +
             
     | 
| 27 | 
         
            +
            ## 📦 Dataset Details
         
     | 
| 28 | 
         
            +
             
     | 
| 29 | 
         
            +
            - **Total images**: 4,114
         
     | 
| 30 | 
         
            +
            - **Annotations**: COCO-style (bounding boxes, human keypoints, both in and out-of-image)
         
     | 
| 31 | 
         
            +
            - **Resolution**: Varies
         
     | 
| 32 | 
         
            +
            - **Format**: JSON annotations + JPG images
         
     | 
| 33 | 
         
            +
             
     | 
| 34 | 
         
            +
            ## 🔍 Tasks
         
     | 
| 35 | 
         
            +
             
     | 
| 36 | 
         
            +
            - **Human Pose Estimation**
         
     | 
| 37 | 
         
            +
            - **Person Detection**
         
     | 
| 38 | 
         
            +
            - **Amodal Person Detection**
         
     | 
| 39 | 
         
            +
             
     | 
| 40 | 
         
            +
            ## 📄 Citation
         
     | 
| 41 | 
         
            +
             
     | 
| 42 | 
         
            +
            If you use CropCOCO in your research, please cite:
         
     | 
| 43 | 
         
            +
             
     | 
| 44 | 
         
            +
            ```bibtex
         
     | 
| 45 | 
         
            +
            @inproceedings{probpose2025,
         
     | 
| 46 | 
         
            +
              title={{ProbPose: A Probabilistic Approach to 2D Human Pose Estimation}}, 
         
     | 
| 47 | 
         
            +
              author={Miroslav Purkrabek and Jiri Matas},
         
     | 
| 48 | 
         
            +
              year={2025},
         
     | 
| 49 | 
         
            +
              booktitle={Computer Vision and Pattern Recognition (CVPR)},
         
     | 
| 50 | 
         
            +
            }
         
     | 
| 51 | 
         
            +
            ```
         
     | 
| 52 | 
         
            +
             
     | 
| 53 | 
         
            +
            ## 🧑💻 Authors
         
     | 
| 54 | 
         
            +
             
     | 
| 55 | 
         
            +
            - Miroslav Purkrabek ([personal website](https://github.com/MiraPurkrabek))
         
     | 
| 56 | 
         
            +
            - Jiri Matas ([personal website](https://cmp.felk.cvut.cz/~matas/))
         
     | 
| 57 | 
         
            +
             
     |