Deep learning for computational pathology
Submitted by Neeraj Kumar (@neerajkumar89) on Monday, 6 June 2016
We strongly believe that the future of not only medical detection and diagnosis but also prognosis and treatment planning will be strongly influenced by pattern recognition and data analysis. Medical imaging will be no different, especially with the advent of techniques such as unsupervised feature extraction and deep learning aided by high performance computing (HPC) in the form of cloud clusters and GPU-based desktops. Currently, we are actively working on pattern recognition applications to histological images. Specifically, we have developed state-of-the art deep learning algorithms for nuclei and mitosis detection, epithelium vs. stroma classification, nuclear abnormality detection etc. In this talk, we will discuss about some of these algorithms and their role in deriving biological insights that can pave the way for improving our understanding of human carcinogenesis.
~Introduction to deep learning and computational pathology (C-path) ~Improtant problems in C-path such as prostate cancer recurrrence prediction that can be addressed using machine learning ~Our deep learning based C-path pipeline ~Important results and future directions
Neeraj Kumar is a research scholar with the department of electronics and electrical engineering at IIT Guwahati. He has developed learning based algorithms for inverse problems such as single image super resolution and reducing the solution space of Non-negative matrix factorization during his PhD. Towards the end of his PhD he shifted the focus of his research to computational pathology and developed deep learning based pipelines for automated analysis of histopathological images. For this purpose he has interned for six months at the Beckman Institute and Department of Bioengineering of the University of Illinois at Urban-Champaign. He is the recipient of the prestigious Microsoft research India fellowship and Erasmus Mundus Heritage Fellowship during his PhD.