SERV-CT数据集是由伦敦大学学院（UCL）外科与介入科学中心（WEISS）开发的手术内窥镜三维重建验证数据集。该数据集包含两组猪组织样本，共计16组立体图像对。每个样本提供完整的相机内外参标定参数、深度图、视差图以及遮挡区域标注。该数据集针对手术内窥镜场景中三维重建算法的验证挑战，特别是在角点特征极少、高反光表面以及存在血液和烟雾的场景下。

内窥镜

└── SERV-CT
    ├── Experiment_1               - frames 001-008
    │   ├── Ground_truth_CT        - The reference data using an O-arm CT scan
    │   │   ├── DepthL             - Left image depth maps (mm depth for each pixel scaled by 256 and stored as 16 bit PNG)
    │   │   ├── DepthR             - Right image depth maps (mm depth for each pixel scaled by 256 and stored as 16 bit PNG)
    │   │   ├── Disparity          - left-to-right disparity (pixel disparity scaled by 256 for subpixel accuracy and stored as 16 bit PNG)
    │   │   ├── OcclusionL         - Colour coded occlusion images
    │   │   └── OcclusionR           (Yellow - non overlap, blue - outside the reference surface, red - not visible in the right image) 
    │   ├── Left_rectified         - left rectified images (720x576 24-bit colour PNG)
    │   ├── Rectified_calibration  - JSON calibration files containing P1, P2 and Q, (units are in pixels and mm)
    │   └── Right_rectified        - right rectified images (720x576 24-bit colour PNG)
    └── Experiment_2               - frames 009-016
        ├── Ground_truth_CT
        │   ├── DepthL
        │   ├── DepthR   
        │   ├── Disparity
        │   ├── OcclusionL
        │   └── OcclusionR
        ├── Ground_truth_RGB       - The reference data using a Creaform RGB scan 
        │   ├── DepthL
        │   ├── DepthR
        │   ├── Disparity
        │   ├── OcclusionL
        │   └── OcclusionR

@article{edwards2020serv,
  title={SERV-CT: A disparity dataset from CT for validation of endoscopic 3D reconstruction},
  author={Edwards, PJ and Psychogyios, Dimitris and Speidel, Stefanie and Maier-Hein, Lena and Stoyanov, Danail},
  journal={arXiv preprint arXiv:2012.11779},
  year={2020}
}

官方网站：
访问官网

下载链接：

百度网盘：

相关论文：
查看论文

发布日期： 2020-12