AbdomenAtlas: A large-scale, detailed-annotated, & multi-center dataset for efficient transfer learning and open algorithmic benchmarking

Wenxuan Li, Chongyu Qu, Xiaoxi Chen, Pedro R. A. S. Bassi, Yijia Shi, Yuxiang Lai, Qian Yu, Huimin Xue, Yixiong Chen, Xiaorui Lin, Yutong Tang, Yining Cao, Haoqi Han, Zheyuan Zhang, Jiawei Liu, Tiezheng Zhang, Yujiu Ma, Jincheng Wang, Guang Zhang, Alan L. Yuille et al. (1 additional authors not shown)

Abstract: Highlights•AbdomenAtlas: The largest abdominal CT dataset of 20,460 per-voxel annotated CT volumes from 112 hospitals.•SuPreM: A suite of pre-trained 3D models enabling efficient transfer learning and few-shot learning.•BodyMaps: An open algorithmic benchmark offering 9K and 5K annotated CT volumes for training and testing.•We develop a semi-automatic annotation procedure for constructing large-scale medical imaging datasets.