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 】

12864_2015_2129_Article_IEq8.gif

12888_2016_877_Article_IEq20.gif

12888_2016_877_Article_IEq19.gif

12711_2017_365_Article_IEq33.gif

12864_2015_2129_Article_IEq5.gif

12888_2016_877_Article_IEq18.gif

12888_2016_877_Article_IEq17.gif

12888_2016_877_Article_IEq16.gif

12888_2016_877_Article_IEq15.gif

12864_2017_3771_Article_IEq6.gif

12864_2017_3687_Article_IEq1.gif

12864_2017_3990_Article_IEq2.gif

12864_2017_4179_Article_IEq26.gif

12864_2016_2793_Article_IEq21.gif

12864_2015_2273_Article_IEq14.gif

12864_2017_4179_Article_IEq24.gif

12864_2017_3715_Article_IEq1.gif

12864_2017_4132_Article_IEq11.gif

12864_2017_4228_Article_IEq2.gif

12864_2017_4132_Article_IEq10.gif

12864_2017_4226_Article_IEq2.gif

12864_2016_2793_Article_IEq13.gif

12864_2017_4030_Article_IEq28.gif

12864_2017_4004_Article_IEq15.gif

12864_2016_2392_Article_IEq2.gif

12864_2017_4004_Article_IEq14.gif

12898_2017_155_Article_IEq33.gif

12864_2017_3733_Article_IEq19.gif

12864_2017_3487_Article_IEq16.gif

12864_2017_4004_Article_IEq12.gif

12864_2017_4004_Article_IEq10.gif

12864_2017_4030_Article_IEq22.gif

12864_2017_4030_Article_IEq20.gif

12864_2017_4186_Article_IEq24.gif