The CP-CHILD dataset is a medical image classification dataset jointly developed by Changsha University of Science and Technology and Hunan Children's Hospital, primarily aimed at machine learning research in gastroenterology. The dataset includes two subsets, CP-CHILD-A and CP-CHILD-B, comprising a total of approximately 9,500 endoscopic images. Each subset is divided into a training set and a test set and is annotated for binary classification based on the presence or absence of polyps.
The primary goal of this dataset is to support the development and evaluation of colonic polyp detection algorithms, which are crucial for the early prevention of colorectal cancer. The dataset adopts a well-defined binary classification structure, where each image is clearly labeled as either polyp or non-polyp, making it particularly suitable for training and evaluating supervised learning models.
Dimensions | Modality | Task Type | Anatomical Structures | Anatomical Area | Number of Categories | Data Volume | File Format |
---|---|---|---|---|---|---|---|
2D | Endoscopy | Classification | Colon | Pelvic | 2 | 9500 | .JPG |
Dataset Statistics | size |
---|---|
min | (256, 256) |
median | (256, 256) |
max | (256, 256) |
Anatomical Structure | Polyp | Non-Polyp |
---|---|---|
Occurrences | 8100 | 1400 |
Percentage | 85.26% | 14.74% |
Minimum Volume (cm³) | 256 x 256 | 256 x 256 |
Median Volume (cm³) | 256 x 256 | 256 x 256 |
Maximum Volume (cm³) | 256 x 256 | 256 x 256 |
No polyp/with polyp comparison.
CP-CHILD
├── CP-CHILD-A
│ ├── Test
│ │ ├── Non-Polyp
│ │ │ ├── 0 (1).jpg
│ │ │ ├── 0 (10).jpg
│ │ │ └── ...
│ │ └── Polyp
│ │ ├── 1 (1).jpg
│ │ ├── 1 (10).jpg
│ │ └── ...
│ └── Train
│ ├── Non-Polyp
│ │ ├── 0 (1).jpg
│ │ ├── 0 (10).jpg
│ │ └── ...
│ └── Polyp
│ ├── 1 (1).jpg
│ ├── 1 (10).jpg
│ └── ...
├── CP-CHILD-B
│ ├── Test
│ │ ├── Non-Polyp
│ │ │ ├── 0 (1).jpg
│ │ │ ├── 0 (10).jpg
│ │ │ └── ...
│ │ └── Polyp
│ │ ├── 1 (1).jpg
│ │ ├── 1 (10).jpg
│ │ └── ...
│ └── Train
│ ├── Non-Polyp
│ │ ├── 0 (1).jpg
│ │ ├── 0 (10).jpg
│ │ └── ...
│ └── Polyp
│ ├── 1 (1).jpg
│ ├── 1 (10).jpg
│ └── ...
└── README.txt
- Wei Wang (Changsha University of Science and Technology)
- Jinge Tian (Changsha University of Science and Technology)
- Chengwen Zhang (Changsha University of Science and Technology)
- Yanhong Luo (Hunan Children's Hospital)
- Xin Wang (Changsha University of Science and Technology)
- Ji Li (Changsha University of Science and Technology)
Official Website:
- https://figshare.com/articles/dataset/CP-CHILD_a_three-class_colon_polyp_dataset/13159811
- kaggle: https://www.kaggle.com/datasets/mah
Download Link: https://www.kaggle.com/datasets/mahdieh002/colonoscopy-cp-child?resource=download
Article Address: https://link.springer.com/article/10.1186/s12880-020-00482-3
Publication Date: 2020-07
@article{wang2020improved,
title={An improved deep learning approach and its applications on colonic polyp images detection},
author={Wang, Wei and Tian, Jinge and Zhang, Chengwen and Luo, Yanhong and Wang, Xin and Li, Ji},
journal={BMC Medical Imaging},
volume={20},
pages={1--14},
year={2020},
publisher={Springer}
}
Original introduction article is here.