Skip to content

Latest commit

 

History

History
121 lines (90 loc) · 4.85 KB

C3VD.md

File metadata and controls

121 lines (90 loc) · 4.85 KB

C3VD Colonoscopy 3D Reconstruction Dataset

Dataset Information

The C3VD (Colonoscopy3DVideoDataset) dataset, developed by Johns Hopkins University, consists of 3D video data from colonoscopies using high-definition clinical colonoscopes and high-fidelity colon models. It serves as a testing benchmark for colonoscopy. The dataset includes 10,015 annotated images, 22 registered videos, and accompanying true depth, surface normals, optical flow, occlusions, six-degree-of-freedom poses, coverage maps, and 3D models. In this work, the authors initially used GANs to generate depth maps of colonoscopy videos, then jointly optimized the camera pose, depth maps, and model rendering results to obtain high-fidelity depth maps, normals, optical flow, and other data, making it the most informative endoscopic dataset available.

The C3VD dataset is of significant importance for the application of 3D computer vision techniques in colonoscopy examinations. It provides an experimental benchmark for evaluating and comparing various computer vision methods, especially in depth estimation, surface reconstruction, and missing area detection. This dataset, through innovative multimodal 2D-3D registration technology, overcomes these challenges, greatly advancing scientific research and technological progress in the field, potentially enhancing the accuracy and efficiency of colonoscopy examinations.

Dataset Meta Information

Dimensions Modality Task Type Anatomical Structures Number of Categories Data Volume File Format
2D Endoscopy 3D Reconstruction Colon 22 10015 png, tiff, txt

Resolution Details

Dataset Statistics size
min 1350 * 1080
median 1350 * 1080
max 1350 * 1080

Label Information Statistics

Section of Colon Number of Images
Cecum 276
Cecum 765
Cecum 370
Cecum 1,142
Cecum 595
Cecum 730
Cecum 465
Cecum 425
Descending Colon 148
Sigmoid Colon 700
Sigmoid Colon 514
Sigmoid Colon 613
Sigmoid Colon 536
Transcending Colon 61
Transcending Colon 700
Transcending Colon 194
Transcending Colon 103
Transcending Colon 235
Transcending Colon 250
Transcending Colon 214
Transcending Colon 382
Transcending Colon 597

Visualization

ITK-SNAP Visualization.

File Structure

Dataset
│
├── cecum_t1_a
│   ├── 0000_color.png
│   ├── 0000_depth.tiff
│   ├── 0000_normals.tiff
│   ├── 0000_occlusion.png
│   └── ...
├── cecum_t1_b
│   ├── 0000_color.png
│   ├── 0000_depth.tiff
│   ├── 0000_normals.tiff
│   ├── 0000_occlusion.png
│   └── ...
└── ...

Authors and Institutions

Taylor L. Bobrow (Johns Hopkins University)

Mayank Golhar (Johns Hopkins University)

Rohan Vijayan (Johns Hopkins University)

Venkata S. Akshintala (Johns Hopkins University)

Juan R. Garcia (Johns Hopkins University)

Nicholas J. Durr (Johns Hopkins University)

Source Information

Official Website: https://durrlab.github.io/C3VD/

Download Link: https://durrlab.github.io/C3VD/

Article Address: https://www.sciencedirect.com/science/article/abs/pii/S1361841523002165

Publication Date: 2023-08

Citation

 @article{bobrow2023,
  title={Colonoscopy 3D video dataset with paired depth from 2D-3D registration},
  author={Bobrow, Taylor L and Golhar, Mayank and Vijayan, Rohan and Akshintala, Venkata S and Garcia, Juan R and Durr, Nicholas J},
  journal={Medical Image Analysis},
  pages={102956},
  year={2023},
  publisher={Elsevier},
}

Original introduction article is here.