Building 99 / 3223
As of Nov. 2011, I joined Rick Szeliski's Interactive Visual Media group at MSR Redmond. Previously, I was a postdoc at the Computation Vision lab at Caltech under Pietro Perona and before that a PhD under the guidance of Serge Belongie at UCSD. I successfully defended my PhD on "Learning from Local Image Regions" in August 2007 and have continued doing research in computer vision and machine learning ever since!
I am interested in machine learning and pattern recognition and their application to computer vision. In particular, I aim to answer two fundamental questions: What are the most effective representations of images and videos that make subsequent reasoning about their content tractable? And what learning paradigms optimally capture the structure and properties of visual information? Understanding the interplay between representation and learning is central toward achieving a more complete understanding of machine vision.
For more information, including project descriptions, code, data, etc., please check out my real webpage!