Towards Large-Scale Histopathological Image Analysis: Hashing-Based Image Retrieval

Tuesday, 11 November 2014 - 2:00pm - 3:00pm
Dr. Shaoting Zhang, Assistant Professor, Department of Computer Science, University of North Carolina at Charlotte
SL 050 IUPUI Campus

Dr. Shaoting Zhang is an Assistant Professor in the Department of Computer Science at the University of North Carolina at Charlotte. Before joining UNC Charlotte, he was a faculty member in the Department of Computer Science at Rutgers-New Brunswick (Research Assistant Professor, 2012-2013). He received his PhD in Computer Science from Rutgers in 01/2012, M.S. from Shanghai Jiao Tong University in 2007, and B.E. from Zhejiang University in 2005. Dr. Zhang's research is on the interface of medical imaging informatics, large-scale visual understanding and machine learning. He has published over 70 articles and registered 6 patents or invention disclosures, including an approved US patent.

His work related to robust and scalable image data analytics has earned multiple awards in top conferences and journals, including the Young Scientist Award (YSA) in MICCAI 2010, the YSA Finalist in MICCAI 2011, one of the top hottest articles (2012, 2013) and most cited articles in Medical Image Analysis, the Front Cover in Medical Physics, May 2013, and the Best Paper Travel Award (3 out of ~600) from ISBI 2014 for the work on mining histopathological images. Dr. Zhang is also a recipient of the 2014 Ralph E. Powe Junior Faculty Enhancement Award from Oak Ridge Associated Universities.


With the ever-increasing amount of annotated medical data, large-scale, data-driven methods provide the promise of bridging the semantic gap between images and diagnoses. The goal of our research project is to increase the scale at which interactive systems can be effective for knowledge discovery in potentially massive database of medical images. Particularly, we focus on the automatic analysis of histopathological images, and propose a scalable image retrieval framework with high-dimensional features extracted in cell-level.

We present a kernelized and supervised hashing method to bridge the semantic gap. With a small amount of supervised information, our method can compress a 10,000-dimensional image feature vector into only tens of binary bits with informative signatures preserved, and these binary codes are then indexed into a hash table that enables real-time retrieval. We validate the hashing-based image retrieval framework on several thousands of images of breast and lung microscopic tissues for both image classification and retrieval. Our framework achieves high search accuracy and promising computational efficiency, comparing favorably with other commonly used methods.

View event flyer