Weakly Supervised Volumetric Segmentation via Self-taught Shape Denoising Model

doi:---

	Weakly Supervised Volumetric Segmentation via Self-taught Shape Denoising Model
	He Q(贺倩)1,2,3 ; Li SL(李帅霖)1 ; He XM(何旭明)1,4
	2021-09
会议录名称	IN MEDICAL IMAGING WITH DEEP LEARNING 2021
卷号	143
页码	268-285
发表状态	已发表
DOI	---
摘要	Weakly supervised segmentation is an important problem in medical image analysis due to the high cost of pixelwise annotation. Prior methods, while often focusing on weak labels of 2D images, exploit few structural cues of volumetric medical images. To address this, we propose a novel weakly-supervised segmentation strategy capable of better capturing 3D shape prior in both model prediction and learning. Our main idea is to extract a self-taught shape representation by leveraging weak labels, and then integrate this representation into segmentation prediction for shape refinement. To this end, we design a deep network consisting of a segmentation module and a shape denoising module, which are trained by an iterative learning strategy. Moreover, we introduce a weak annotation scheme with a hybrid label design for volumetric images, which improves model learning without increasing the overall annotation cost. The empirical experiments show that our approach outperforms existing SOTA strategies on three organ segmentation benchmarks with distinctive shape properties. Notably, we can achieve strong performance with even 10% labeled slices, which is significantly superior to other methods. Our code is available at: https://github.com/Seolen/weak seg via shape model. © 2021 Q. He, S. Li & X. He.
关键词	Cost benefit analysis Image enhancement Image segmentation Iterative methods Medical imaging 3-D shape 3d shape prior De-noising Model learning Self-taught learning Shape priors Supervised segmentation Volumetric segmentations Weak labels Weakly supervised segmentation
会议名称	4th Conference on Medical Imaging with Deep Learning, MIDL 2021
会议地点	Virtual, Online, Germany
会议日期	July 7, 2021 - July 9, 2021
收录类别	SCI ; EI
语种	英语
出版者	ML Research Press
EI入藏号	20232614305227
EI主题词	Learning systems
EISSN	2640-3498
EI分类号	461.1 Biomedical Engineering ; 746 Imaging Techniques ; 911 Cost and Value Engineering ; Industrial Economics ; 912.2 Management ; 921.6 Numerical Methods
原始文献类型	Conference article (CA)
文献类型	会议论文
条目标识符	https://kms.shanghaitech.edu.cn/handle/2MSLDSTB/128385
专题	信息科学与技术学院_博士生信息科学与技术学院_PI研究组_何旭明组信息科学与技术学院_硕士生
共同第一作者	Li SL(李帅霖)
通讯作者	He XM(何旭明)
作者单位	1.School of Information Science and Technology, ShanghaiTech University 2.Shanghai Institute of Microsystem and Information Technology, Chinese Academy of Sciences 3.University of Chinese Academy of Sciences 4.Shanghai Engineering Research Center of Intelligent Vision and Imaging
第一作者单位	信息科学与技术学院
通讯作者单位	信息科学与技术学院
第一作者的第一单位	信息科学与技术学院
推荐引用方式 GB/T 7714	He Q,Li SL,He XM. Weakly Supervised Volumetric Segmentation via Self-taught Shape Denoising Model[C]:ML Research Press,2021:268-285.