Computer Vision and Machine Learning for Cancerous Tissue Recognition

Nikos Papanikolopoulos
McKnight Presidential Endowed Professor, Distinguished McKnight University Professor
University of Minnesota

4:10pm Monday, January 29, 2018
Room 124 HRBB


Today, vast and unwieldy data collections are regularly being generated and analyzed in hopes of supporting an ever-expanding range of challenging sensing applications. Modern inference schemes usually involve millions of parameters to learn complex real-world tasks, which creates the need for large annotated datasets for training. For several visual learning applications, collecting large amounts of annotated data is either challenging or very expensive; one such domain is medical image analysis. In this work, machine learning methods were devised with emphasis on Cancerous Tissue Recognition (CTR) applications.

First, a lightweight active constrained clustering scheme was developed for the processing of image data which capitalizes on actively acquired pairwise constraints. The proposed methodology introduces the use of the Silhouette values, conventionally used for measuring clustering performance, in order to rank the degree of information content of the various samples. Second, an active selection framework that operates in tandem with Convolutional Neural Networks (CNNs) was constructed for CTR. In the presence of limited annotations, alternative (or sometimes complementary) venues were explored in an effort to restrain the high expenditure of collecting image annotations required by CNN-based schemes.

Third, a Symmetric Positive Definite (SPD) image representation was derived for CTR, termed Covariance Kernel Descriptor (CKD) which consistently outperformed a large collection of popular image descriptors. Even though the CKD successfully describes the tissue architecture for small image regions, its performance decays when implemented on larger slide regions or whole tissue slides due to the larger variability that tissue exhibits at that level, since different types of tissue can be present as the regions grow (healthy, benign disease, malignant disease). Fourth, to leverage the recognition capability of the CKDs to larger slide regions, the Weakly Annotated Image Descriptor (WAID) was devised as the parameters of classifier decision boundaries in a multiple instance learning framework.

*This is joint work with Panos Stanitsas and Sasha Truskinovsky


Nikolaos P. Papanikolopoulos (IEEE Fellow) received the Diploma degree in electrical and computer engineering from the National Technical University of Athens, Athens, Greece, in 1987, the M.S.E.E. in electrical engineering from Carnegie Mellon University (CMU), Pittsburgh, PA, in 1988, and the Ph.D. in electrical and computer engineering from Carnegie Mellon University, Pittsburgh, PA, in 1992.  Currently, he is a McKnight Presidential Endowed Professor and a Distinguished McKnight University Professor in the Department of Computer Science at the University of Minnesota and Director of the Center for Distributed Robotics and SECTTRA. His research interests include robotics, computer vision, sensors for transportation applications,  and control. He has authored or co-authored more than 350 journal and conference papers in the above areas (more than seventy five refereed journal papers).

Faculty Contact: Dr. Nancy M. Amato

Building Concepts Bridges: Knowledge Discovery from Literature and Beyond

Aidong Zhang
SUNY Distinguished Professor
University at Buffalo

4:10pm Monday, February 19, 2017
Room 124 HRBB


With the growth of world wide web and large-scale digitization of documents, we are overwhelmed with massive information, formally through publication of various scientific journals or informally through internet. As an example, consider MEDLINE, a premier bibliographic database in life sciences, with currently more than 23 million references from approximately 5,600 worldwide journals. As a consequence, Literature Based Discovery has become a sub-field of Text Mining that leverages these published articles to formulate hypotheses. In this talk, I will discuss how a self-learning based framework for knowledge discovery can be designed to mine hidden associations between non-interacting scientific concepts by rationally connecting independent nuggets of published literature. The self-learning process can model the evolutionary behavior of concepts to uncover latent associations between text concepts, which allows us to learn the evolutionary trajectories of text terms and detect informative terms in a completely unsupervised manner. Hence, meaningful hypotheses can be efficiently generated without prior knowledge. I will also discuss how this self-learning framework can be extended to include social media and Internet forums. With the capability to discern reliable information from various sources, this self-learning framework provides a platform for combining heterogeneous sources and intelligently learning new knowledge with no user intervention.


Dr. Aidong Zhang is a SUNY Distinguished Professor of Computer Science and Engineering at the State University of New York (SUNY) at Buffalo where she served as Department Chair from 2009 to 2015. She is currently on leave and serving as Program Director in the Information & Intelligent Systems Division of the Directorate for Computer & Information Science & Engineering, National Science Foundation. Her research interests include data analytics/data science, machine learning, bioinformatics, and health informatics, and she has authored over 300 research publications in these areas. Dr. Zhang currently serves as the Editor-in-Chief of the IEEE Transactions on Computational Biology and Bioinformatics (TCBB). She served as the founding Chair of ACM Special Interest Group on Bioinformatics, Computational Biology and Biomedical Informatics during 2011-2015 and is currently Chair of its advisory board. She is also the founding and steering chair of ACM international conference on Bioinformatics, Computational Biology and Health Informatics. She has served as editor for several other journal editorial boards, and has also chaired or served on numerous program committees of international conferences and workshops. Dr. Zhang is both an ACM Fellow and an IEEE Fellow.

Faculty Contact: Dr. Xia (Ben) Hu