Monday, November 16, 2015 - 4:00pm to 4:50pm
KEC 1003

Speaker Information

Fuxin Li
Assistant Professor
School of EECS
Oregon State University

Abstract

Many of the structural prediction problems in computer vision can be converted into simple regression problems on object proposals followed by specific higher-order inference schemes. The benefit of this approach is that learning can be simple without the need of doing structural prediction, while inference carries the bulk of the optimization load. This approach started from object proposals in semantic segmentation, while carrying over to many other problems such as video object discovery and multi-target tracking, where the efficiency of the least squares regression helps learning thousands of models with ease. A new type of higher order inference called composite statistical inference is proposed to utilize those regression estimates. This inference breaks and recombines object proposals and is flexible, higher-order, and statistically consistent. It is shown to work well in complex problems with significant interactions among objects.

Speaker Bio

Before coming to OSU, Dr. Fuxin Li was a research scientist from the School of Interactive Computing, Georgia Institute of Technology, working with Dr. James M. Rehg. He obtained a Ph.D. degree in 2009 from the Institute of Automation, Chinese Academy of Sciences and had since held postdoctoral appointments in the University of Bonn and Georgia Institute of Technology. Dr. Li is interested in the intersection of machine learning and computer vision, especially segmentation in images/videos and visual object recognition. He has published more than 30 papers, including more than 15 papers in top conferences in machine learning and computer vision. He and his colleagues have won the prestigious PASCAL Visual Object Recognition challenge in Segmentation from 2009-2012. He has won a Microsoft research award and 2 best reviewer awards.