All-Around Real Label Supervision: Cyclic Prototype Consistency Learning for Semi-Supervised Medical Image Segmentation

doi:10.1109/JBHI.2022.3162043

	All-Around Real Label Supervision: Cyclic Prototype Consistency Learning for Semi-Supervised Medical Image Segmentation
	Xu, Zhe 1; Wang, Yixin 4; Lu, Donghuan 2,3; Yu, Lequan 5; Yan, Jiangpeng 6; Luo, Jie 7; Ma, Kai 2,3; Zheng, Yefeng 2,3; Tong, Raymond Kai-yu 1
	2022-07-01
发表期刊	IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS
ISSN	2168-2194
卷号	26 期号:7 页码:3174-3184
摘要	Semi-supervised learning has substantially advanced medical image segmentation since it alleviates the heavy burden of acquiring the costly expert-examined annotations. Especially, the consistency-based approaches have attracted more attention for their superior performance, wherein the real labels are only utilized to supervise their paired images via supervised loss while the unlabeled images are exploited by enforcing the perturbation-based "unsupervised" consistency without explicit guidance from those real labels. However, intuitively, the expert-examined real labels contain more reliable supervision signals. Observing this, we ask an unexplored but interesting question: can we exploit the unlabeled data via explicit real label supervision for semi-supervised training? To this end, we discard the previous perturbation-based consistency but absorb the essence of non-parametric prototype learning. Based on the prototypical networks, we then propose a novel cyclic prototype consistency learning (CPCL) framework, which is constructed by a labeled-to-unlabeled (L2U) prototypical forward process and an unlabeled-to-labeled (U2L) backward process. Such two processes synergistically enhance the segmentation network by encouraging morediscriminative and compact features. In this way, our framework turns previous "unsupervised" consistency into new "supervised" consistency, obtaining the "all-around real label supervision" property of our method. Extensive experiments on brain tumor segmentation from MRI and kidney segmentation from CT images show that our CPCL can effectively exploit the unlabeled data and outperform other state-of-the-art semi-supervised medical image segmentation methods.
关键词	Image segmentation Prototypes Biomedical imaging Perturbation methods Reliability Feature extraction Training Medical image segmentation prototype learning semi-supervised learning
DOI	10.1109/JBHI.2022.3162043
收录类别	SCI
语种	英语
资助项目	National Key R&D Program of China[2018YFC2000702] ; General Research Fund from Research Grant Council of Hong Kong[14205419] ; Tencent Healthcare (Shenzhen) Co., LTD ; Tencent Jarvis Lab
WOS研究方向	Computer Science ; Mathematical & Computational Biology ; Medical Informatics
WOS类目	Computer Science, Information Systems ; Computer Science, Interdisciplinary Applications ; Mathematical & Computational Biology ; Medical Informatics
WOS记录号	WOS:000819832600033
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
引用统计	被引频次：36[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/19511
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Lu, Donghuan; Tong, Raymond Kai-yu
作者单位	1.Chinese Univ Hong Kong, Dept Biomed Engn, Hong Kong 999077, Peoples R China 2.Tencent Hlthcare Shenzhen Co LTD, Shenzhen 518000, Peoples R China 3.Tencent Jarvis Lab, Shenzhen 518000, Peoples R China 4.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 5.Univ Hong Kong, Dept Stat & Actuarial Sci, Hong Kong, Peoples R China 6.Tsinghua Univ, Dept Automat, Beijing 100084, Peoples R China 7.Harvard Med Sch, Brigham & Womens Hosp, Boston, MA 02115 USA
推荐引用方式 GB/T 7714	Xu, Zhe,Wang, Yixin,Lu, Donghuan,et al. All-Around Real Label Supervision: Cyclic Prototype Consistency Learning for Semi-Supervised Medical Image Segmentation[J]. IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS,2022,26(7):3174-3184.
APA	Xu, Zhe.,Wang, Yixin.,Lu, Donghuan.,Yu, Lequan.,Yan, Jiangpeng.,...&Tong, Raymond Kai-yu.(2022).All-Around Real Label Supervision: Cyclic Prototype Consistency Learning for Semi-Supervised Medical Image Segmentation.IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS,26(7),3174-3184.
MLA	Xu, Zhe,et al."All-Around Real Label Supervision: Cyclic Prototype Consistency Learning for Semi-Supervised Medical Image Segmentation".IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS 26.7(2022):3174-3184.