Weakly supervised histopathology cancer image segmentation and classification
People
Abstract
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.
Paper
CVPR 2012 paper, 2.0MB
MIA 2014 paper, 7.6MB
Citation
Yan Xu*, Jun-Yan Zhu*, Eric Chang and Zhuowen Tu. (*equal contribution). "Multiple Clustered Instance Learning for Histopathology Cancer Image Segmentation, Clustering, and Classification", in IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2012. Bibtex
Yan Xu, Jun-Yan Zhu, Eric I-Chao Chang, Maode Lai and Zhuowen Tu. "Weakly Supervised Histopathology Cancer Image Segmentation and Classification", in Medical Image Analysis (MIA), 2014. Bibtex
Software
-
MCILBoost: A boosting-based Multiple Instance Learning (MIL) package that includes MIL-Boost [Viola et al. 2006] and MCIL-Boost. The core of this package is a command line interface written in C++. Various MATLAB helper functions are provided to help users easily train/test MCIL-Boost model, perform cross-validation, and evaluate the performance.
Additional Materials
Acknowledgement
We would like to thank Department of Pathology, Zhejiang University in China for providing data and help.
Related Papers
Yan Xu, Jianwen Zhang, Eric Chang, Maode Lai, and Zhuowen Tu. "Contexts-Constrained Multiple Instance Learning for Histopathology Image Segmentation", in International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), 2012.
Funding
This research is supported in part by:
- Microsoft Research Asia eHealth grant
- NSF CAREER award IIS-0844566 (IIS-1360568)
- NSF IIS-1216528 (IIS-1360566)
- ONR N000140910099
- Grant 61073077 from National Science Foundation of China
- Grant SKLSDE-2011ZX-13 from Beihang University in China
