EE Colloquium: Domain Enriched Learning for Image Processing and Computer Vision

Abstract: This talk will survey current research activity in the Information Processing and Algorithms (iPAL) Lab at Penn State. A theme that resonates throughout the talk is the integration of problem specific domain knowledge and physical models with learning frameworks such as modern deep neural networks. Among the flavors of domain enrichment: 1.) we will discuss the integration of priors on images that enables success in scenarios where training data is limited and/or imperfect, and 2.) algorithm unrolling as a path towards designing interpretable neural network structures that afford superior generalizability. Applications will be discussed in imaging inverse problems as well as classification and recognition.   

Bio: Professor Vishal Monga has been on the EE faculty at Penn State since Fall 2009. From Oct 2005-July 2009 he was an imaging scientist with Xerox Research Labs. He has also been a visiting researcher at Microsoft Research in Redmond, WA and a visiting faculty at the University of Rochester. Prior to that, he received his PhDEE from the department of Electrical and Computer Engineering at the University of Texas, Austin. Dr. Monga is an elected member of the IEEE Image Video and Multidimensional Signal Processing (IVMSP) Technical Committee and has served on many editorial boards including the IEEE Transactions on Image Processing, IEEE Signal Processing Letters and IEEE Transactions on Circuits and Systems for Video Technology. Professor Monga's research has been recognized via the 2015 US National Science Foundation CAREER award and a 2019 Penn State Engineering Alumni Society (PSEAS) Outstanding Research Award. For his educational efforts, Dr. Monga received the 2016 Joel and Ruth Spira Teaching Excellence award. His group’s work focuses on convex and non-convex optimization based methods with applications in learning, vision and signal processing.