Parallel multiple instance learning for extremely large histopathology image analysis

BMC Bioinformatics

Research Article

Ziwei Wu¹  Maode Lai²  Teng Gao³  Eric I-Chao Chang³  Yeshu Li⁴  Yubo Fan⁵  Zhengyang Shen⁵  Yan Xu⁶ 

[1] College of Information and Electrical Engineering, China Agricultural University, Beijing, China;Department of Pathology, School of Medicine, Zhejiang University, Zhejiang, China;Microsoft Research Asia, Beijing, China;School of Computer Science and Engineering, Beihang University, Beijing, China;State Key Laboratory of Software Development Environment and Key Laboratory of Biomechanics and Mechanobiology of Ministry of Education and Research Institute of Beihang University in Shenzhen, Beijing, China;State Key Laboratory of Software Development Environment and Key Laboratory of Biomechanics and Mechanobiology of Ministry of Education and Research Institute of Beihang University in Shenzhen, Beijing, China;Microsoft Research Asia, Beijing, China;

关键词: Histopathology image analysis;    Microscopic image analysis;    Multiple instance learning;    Parallelization;   

DOI  :  10.1186/s12859-017-1768-8

 received in 2017-03-08, accepted in 2017-07-19,  发布年份 2017

来源: Springer

【 摘 要 】

BackgroundHistopathology images are critical for medical diagnosis, e.g., cancer and its treatment. A standard histopathology slice can be easily scanned at a high resolution of, say, 200,000×200,000 pixels. These high resolution images can make most existing imaging processing tools infeasible or less effective when operated on a single machine with limited memory, disk space and computing power.ResultsIn this paper, we propose an algorithm tackling this new emerging “big data” problem utilizing parallel computing on High-Performance-Computing (HPC) clusters. Experimental results on a large-scale data set (1318 images at a scale of 10 billion pixels each) demonstrate the efficiency and effectiveness of the proposed algorithm for low-latency real-time applications.ConclusionsThe framework proposed an effective and efficient system for extremely large histopathology image analysis. It is based on the multiple instance learning formulation for weakly-supervised learning for image classification, segmentation and clustering. When a max-margin concept is adopted for different clusters, we obtain further improvement in clustering performance.

【 授权许可】

CC BY   
© The Author(s) 2017

【 图 表 】

12936_2017_2075_Article_IEq66.gif

Fig. 1

Fig. 1

Fig. 3

12936_2017_2045_Article_IEq5.gif

Fig. 5

12936_2017_2045_Article_IEq3.gif

Fig. 2

Fig. 2

12951_2015_155_Article_IEq56.gif

Fig. 3

12951_2017_255_Article_IEq36.gif

Fig. 2

Fig. 2

Fig. 1

Fig. 1

Fig. 2

12951_2015_155_Article_IEq54.gif

12951_2015_155_Article_IEq53.gif

12951_2017_255_Article_IEq34.gif

12951_2017_255_Article_IEq33.gif

12936_2016_1316_Article_IEq8.gif

Fig. 3

Fig. 8

12936_2017_1963_Article_IEq60.gif

Fig. 1

Fig. 3