Weakly supervised histopathology cancer image segmentation and classification

被引:223
作者
Xu, Yan [1 ,2 ]
Zhu, Jun-Yan [3 ]
I-Chao, Eric [2 ]
Lai, Maode [4 ]
Tu, Zhuowen [5 ]
机构
[1] Beihang Univ, Minist Educ, State Key Lab Software Dev Environm, Key Lab Biomech & Mechanobiol, Beijing, Peoples R China
[2] Microsoft Res Asia, Beijing 10080, Peoples R China
[3] Univ Calif Berkeley, Div Comp Sci, Berkeley, CA 94720 USA
[4] Zhejiang Univ, Sch Med, Dept Pathol, Hangzhou, Zhejiang, Peoples R China
[5] Univ Calif San Diego, Dept Cognit Sci, San Diego, CA 92103 USA
基金
美国国家科学基金会;
关键词
Image segmentation; Classification; Clustering; Multiple instance learning; Histopathology image; PROSTATE-CANCER; INSTANCE; NEUROBLASTOMA; FEATURES; WAVELET; SCALE;
D O I
10.1016/j.media.2014.01.010
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Labeling a histopathology image as having cancerous regions or not is a critical task in cancer diagnosis; it is also clinically important to segment the cancer tissues and cluster them into various classes. Existing supervised approaches for image classification and segmentation require detailed manual annotations for the cancer pixels, which are time-consuming to obtain. In this paper, we propose a new learning method, multiple clustered instance learning (MCIL) (along the line of weakly supervised learning) for histopathology image segmentation. The proposed MCIL method simultaneously performs image-level classification (cancer vs. non-cancer image), medical image segmentation (cancer vs. non-cancer tissue), and patch-level clustering (different classes). We embed the clustering concept into the multiple instance learning (MIL) setting and derive a principled solution to performing the above three tasks in an integrated framework. In addition, we introduce contextual constraints as a prior for MCIL, which further reduces the ambiguity in MIL. Experimental results on histopathology colon cancer images and cytology images demonstrate the great advantage of MCIL over the competing methods. (C) 2014 Elsevier B.V. All rights reserved.
引用
收藏
页码:591 / 604
页数:14
相关论文
共 66 条
[1]   Color Graphs for Automated Cancer Diagnosis and Grading [J].
Altunbay, Dogan ;
Cigir, Celal ;
Sokmensuer, Cenk ;
Gunduz-Demir, Cigdem .
IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, 2010, 57 (03) :665-674
[2]  
Andrews S., 2003, NEURAL INFORM PROCES
[3]  
[Anonymous], 1999, Athena scientific Belmont
[4]  
[Anonymous], INT JOINT C ART INT
[5]  
[Anonymous], 2000, ICML WORKSH ATTR VAL
[6]  
[Anonymous], EUR C COMP VIS
[7]  
[Anonymous], INT C MACH LEARN
[8]  
[Anonymous], INT C MACH LEARN
[9]  
[Anonymous], 2000, ADV NEURAL INFORM PR
[10]  
[Anonymous], EUR C COMP VIS