VALSE 论文速览 第198期：AbdomenAtlas-8K

论文题目：

AbdomenAtlas-8K: Annotating 8,000 CT Volumes for Multi-Organ Segmentation in Three Weeks

作者列表：

Chongyu Qu (Johns Hopkins University), Wenxuan Li (Johns Hopkins University), Tiezheng Zhang (Johns Hopkins University), Hualin Qiao (Rutgers University), Jie Liu (City University of Hong Kong), Yucheng Tang (NVIDIA), Alan L. Yuille (Johns Hopkins University), and Zongwei Zhou (Johns Hopkins University)

B站观看网址：

https://www.bilibili.com/video/BV1voxNe8EEB/?spm_id_from=333.999.0.0&vd_source=2044ac986b5caa4c8b5f00b525441df3

复制链接到浏览器打开或点击阅读原文即可跳转至观看页面。

论文摘要：

Annotating medical images, particularly for organ segmentation, is laborious and time-consuming. For example, annotating an abdominal organ requires an estimated rate of 30--60 minutes per CT volume based on the expertise of an annotator and the size, visibility, and complexity of the organ. Therefore, publicly available datasets for multi-organ segmentation are often limited in data size and organ diversity. This paper proposes an active learning procedure to expedite the annotation process for organ segmentation and creates the largest multi-organ dataset (by far) with the spleen, liver, kidneys, stomach, gallbladder, pancreas, aorta, and IVC annotated in 8448 CT volumes, equating to 3.2 million slices. The conventional annotation methods would take an experienced annotator up to 1,600 weeks (or roughly 30.8 years) to complete this task. In contrast, our annotation procedure has accomplished this task in three weeks (based on an 8-hour workday, five days a week) while maintaining a similar or even better annotation quality. This achievement is attributed to three unique properties of our method: (1) label bias reduction using multiple pre-trained segmentation models, (2) effective error detection in the model predictions, and (3) attention guidance for annotators to make corrections on the most salient errors. Furthermore, we summarize the taxonomy of common errors made by AI algorithms and annotators. This allows for continuous improvement of AI and annotations, significantly reducing the annotation costs required to create large-scale datasets for a wider variety of medical imaging tasks.

参考文献：

[1] Chongyu Qu, Tiezheng Zhang, Hualin Qiao, Jie Liu, Yucheng Tang, Alan Yuille. and Zongwei Zhou*., AbdomenAtlas-8K: Annotating 8,000 CT Volumes for Multi-Organ Segmentation in Three Weeks. NeurIPS, 2023

论文链接：

[https://arxiv.org/abs/2305.09666]

代码链接：

[GitHub (AbdomenAtlas 1.0): https://github.com/MrGiovanni/AbdomenAtlas]

[GitHub (AbdomenAtlas 1.1): https://github.com/MrGiovanni/SuPreM]

视频讲者简介：

Chongyu Qu is a research assistant at CCVL (Computational Cognition, Vision, and Learning) Lab at Johns Hopkins University, under the advisement of Prof. Alan L.Yuille and Dr. Zongwei Zhou. He received the M.S.E degree in Biomedical Engineering from Johns Hopkins University in 2022. His research interests lie in the intersection of computer vision and medical image analysis. His current focus is on developing a robust medical foundation model with adaptability to various downstream applications. The ultimate objective is to advance innovations in computer-aided diagnosis. This involves assisting experts in acquiring a more comprehensive understanding of the human body, delivering precise diagnoses, and enabling early detection. 

个人主页：

https://chongyu1117.github.io/website/

特别鸣谢本次论文速览主要组织者：

月度轮值AC：于茜 (北京航空航天大学)