CADS-dataset
收藏CADS数据集概述
数据集基本信息
- 数据集名称:CADS (Comprehensive Anatomical Dataset and Segmentation for Whole-Body Anatomy in Computed Tomography)
- 主要任务:图像分割
- 任务类别:image-segmentation
- 数据模态:医学CT图像
- 图像维度:3D
- 覆盖范围:全身(从头到膝盖区域)
- 数据规模:10K < n < 100K
- 许可协议:cadsdataset (其他)
- 许可协议链接:https://github.com/murong-xu/CADS
数据集核心内容
CADS是一个用于在计算机断层扫描(CT)中分割167个解剖结构的稳健、全自动框架。该框架包含两个主要组件,本仓库托管的是CADS-dataset。
CADS-dataset
- 包含22,022个CT体积,具有167个解剖结构的完整注释。
- 是目前最广泛的全身CT数据集,在规模(比现有集合多18倍CT扫描)和解剖覆盖范围(多60%的不同目标)上均超过当前数据集。
- 数据收集自公开可用数据集和私人医院数据,涵盖16个国家100多个成像中心。
- 覆盖了临床变异性、协议和病理状况的多样性。
- 通过具有伪标记和无监督质量控制的自动化流程构建。
数据格式与结构
-
所有图像和分割结果均以NIfTI格式提供,按数据源组织。
-
目录结构为:
root/ ├── dataset_name/ │ ├── images/ # 原始CT体积 │ ├── segmentations/ # 分割掩码(索引参见模型标签映射) │ └── README.md # 数据集许可、引用和详细信息
数据集配置与来源
CADS-dataset包含多个公开可用和私人来源的数据集,每个数据集均在其自己的许可下发布。共包含43个配置(数据子集),具体如下:
| 配置名称 | 数据集名称 | 许可协议 | CT体积数量 |
|---|---|---|---|
| 0001_visceral_gc | VISCERAL Gold Corpus | Customized license | 40 |
| 0002_visceral_sc | VISCERAL Silver Corpus | Customized license | 127 |
| 0003_kits21 | The Kidney and Kidney Tumor Segmentation Challenge (KiTS21) | CC BY-NC-SA 4.0 | 300 |
| 0004_lits | Liver Tumor Segmentation Benchmark (LiTS) | CC BY-NC-SA 4.0 | 201 |
| 0005_bcv_abdomen | MICCAI Multi-Atlas Labeling Beyond the Cranial Vault (Abdomen) | CC BY 4.0 | 50 |
| 0006_bcv_cervix | MICCAI Multi-Atlas Labeling Beyond the Cranial Vault (Cervix) | CC BY 4.0 | 50 |
| 0007_chaos | CHAOS – Combined (CT-MR) Healthy Abdominal Organ Segmentation Challenge (CT Subset) | CC BY-NC-SA 4.0 | 40 |
| 0008_ctorg | CT-ORG: Multiple Organ Segmentation in CT | CC BY 3.0 | 140 |
| 0009_abdomenct1k | AbdomenCT-1K | CC BY 4.0 | 1062 |
| 0010_verse | VerSe – Vertebrae Labelling and Segmentation Benchmark | CC BY-SA 4.0 | 374 |
| 0011_exact | EXACT09 – Extraction of Airways from CT | Customized license | 40 |
| 0012_cad_pe | CAD-PE – Computer Aided Detection for Pulmonary Embolism Challenge | CC BY 4.0 | 40 |
| 0013_ribfrac | RibFrac Challenge Dataset | CC BY-NC 4.0 | 660 |
| 0014_learn2reg | Learn2Reg – Abdomen MR-CT (TCIA Subset) | CC BY 3.0 and TCIA Data Usage Policy | 16 |
| 0015_lndb | LNDb – Lung Nodule Database | CC BY-NC-ND 4.0 | 294 |
| 0016_lidc | LIDC-IDRI – Lung Image Database Consortium and Image Database Resource Initiative | CC BY 3.0 | 997 |
| 0017_lola11 | LOLA11 (LObe and Lung Analysis 2011) | Customized license | 55 |
| 0018_sliver07 | SLIVER07 (Segmentation of the Liver 2007) | Customized license | 30 |
| 0019_tcia_ct_lymph_nodes | Lymph Node CT Dataset (NIH, TCIA) | CC BY 3.0 | 174 |
| 0020_tcia_cptac_ccrcc | CPTAC-CCRCC – Clear Cell Renal Cell Carcinoma | CC BY 3.0 | 258 |
| 0021_tcia_cptac_luad | CPTAC-LUAD – Clinical Proteomic Tumor Analysis Consortium Lung Adenocarcinoma Collection | CC BY 3.0 | 133 |
| 0022_tcia_ct_images_covid19 | CT Images in COVID-19 | CC BY 4.0 | 121 |
| 0023_tcia_nsclc_radiomics | NSCLC Radiogenomics | CC BY 3.0 | 131 |
| 0024_pancreas_ct | Pancreas-CT | CC BY 3.0 | 80 |
| 0025_pancreatic_ct_cbct_seg | Pancreatic CT-CBCT Segmentation | CC BY 4.0 | 93 |
| 0026_rider_lung_ct | RIDER Lung CT | CC BY 4.0 | 59 |
| 0027_tcia_tcga_kich | TCGA-KICH (Kidney Chromophobe) | CC BY 3.0 | 17 |
| 0028_tcia_tcga_kirc | TCGA-KIRC (Kidney Renal Clear Cell Carcinoma) | CC BY 3.0 | 398 |
| 0029_tcia_tcga_kirp | TCGA-KIRP (Kidney Renal Papillary Cell Carcinoma) | CC BY 3.0 | 19 |
| 0030_tcia_tcga_lihc | TCGA-LIHC (Liver Hepatocellular Carcinoma) | CC BY 3.0 | 242 |
| 0032_stoic2021 | STOIC (Study of Thoracic CT in COVID-19) | CC BY-NC 4.0 | 2000 |
| 0033_tcia_nlst | National Lung Screening Trial (NLST) | CC BY 4.0 | 7172 |
| 0034_empire | EMPIRE10 Challenge | Customized license | 60 |
| 0037_totalsegmentator | TotalSegmentator | CC BY 4.0 | 1203 |
| 0038_amos | AMOS (Multi-Modality Abdominal Multi-Organ Segmentation Challenge) | CC BY 4.0 | 200 |
| 0039_han_seg | HaN-Seg: The head and neck organ-at-risk CT & MR segmentation dataset | CC BY-NC-ND 4.0 | 42 |
| 0040_saros | SAROS: A dataset for whole-body region and organ segmentation in CT imaging | Mix of CC BY 3.0, CC BY 4.0, and CC BY-NC 3.0 | 900 |
| 0041_ctrate | CT-RATE | CC BY-NC-SA 4.0 | 3134 |
| 0042_new_brainct_1mm | (Newly Released) BrainCT-1mm | CC BY 4.0 | 484 |
| 0043_new_ct_tri | (Newly Released) CT-TRI (Triphasic Contrast-Enhanced Abdominal CTs) | CC BY-NC-SA 4.0 | 586 |
重要说明
- 除在本项目中新发布的BrainCT-1mm和CT-TRI数据集外,并非CT图像的原始所有者。
- 用户在使用数据前应查看每个数据集子目录中的相应README.md文件,并根据预期用途决定是否包含或排除该数据集。
- 更新(2025-10-04):修复了数据集
0010_verse、0041_ctrate和0043_new_ct_tri中缺失的图像并校正了仿射/强度错误。
相关资源
- CADS论文预印本:https://arxiv.org/abs/2507.22953
- CADS-dataset:https://huggingface.co/datasets/mrmrx/CADS-dataset
- CADS-model权重:https://github.com/murong-xu/CADS/releases/tag/cads-model_v1.0.0
- CADS-model代码库:https://github.com/murong-xu/CADS
- CADS-model 3D Slicer插件:https://github.com/murong-xu/SlicerCADSWholeBodyCTSeg
引用
如果使用CADS的任何组件(CADS-dataset、其策划的分割掩码、预训练的CADS-model或3D Slicer扩展),请引用: bibtex @article{xu2025cads, title={CADS: A Comprehensive Anatomical Dataset and Segmentation for Whole-Body Anatomy in Computed Tomography}, author={Xu, Murong and Amiranashvili, Tamaz and Navarro, Fernando and Fritsak, Maksym and Hamamci, Ibrahim Ethem and Shit, Suprosanna and Wittmann, Bastian and Er, Sezgin and Christ, Sebastian M. and de la Rosa, Ezequiel and Deseoe, Julian and Graf, Robert and Möller, Hendrik and Sekuboyina, Anjany and Peeken, Jan C. and Becker, Sven and Baldini, Giulia and Haubold, Johannes and Nensa, Felix and Hosch, René and Mirajkar, Nikhil and Khalid, Saad and Zachow, Stefan and Weber, Marc-André and Langs, Georg and Wasserthal, Jakob and Ozdemir, Mehmet Kemal and Fedorov, Andrey and Kikinis, Ron and Tanadini-Lang, Stephanie and Kirschke, Jan S. and Combs, Stephanie E. and Menze, Bjoern}, journal={arXiv preprint arXiv:2507.22953}, year={2025} }




