ICVGIP

Abstract

Many problems in Computer Vision are formulated using Markov models. Solution under this model involves computing the most probable values of certain random variables. This problem, known as Maximum a Posteriori (MAP) estimation has been widely studied in Computer Science and the resulting algorithms have led to accurate and reliable solutions for many problems in computer vision and information engineering. This tutorial is aimed at researchers who wish to use and understand these algorithms. The tutorial will answer the following questions: How to formalize and solve some known vision problems using MAP inference of a random field? What are the different genres of MAP inference algorithms, and how do they work? Which algorithm is suitable for which problem? and lastly, what are the recent developments and open questions in this field.

About the speakers

Pushmeet Kohli
http://research.microsoft.com/en-us/um/people/pkohli/
Machine Learning and Perception
Microsoft Research
Cambridge, UK
pkohli@microsoft.com

Pushmeet Kohli is a researcher in the Machine Learning and Perception group at Microsoft Research Cambridge, and a post-doctoral associate of Trinity Hall, University of Cambridge. He completed his PhD studies at Oxford Brookes University under the supervision of Prof. Philip Torr. His PhD thesis, titled "Minimizing Dynamic and Higher Order Energy Functions using Graph Cuts", was the winner of the British Machine Vision Association's Sullivan Doctoral Thesis Award, and was a runner-up for the British Computer Society's Distinguished Dissertation Award. Before joining Microsoft Research Cambridge, Pushmeet was a visiting researcher at Microsoft Research Bangalore. He previously worked in the Foundation of Software Engineering Group at MSR Redmond. Pushmeet has worked on a number of problems in Computer Vision, Machine Learning and Discrete Optimization. His papers have appeared in SIGGRAPH, PAMI, IJCV, ICCV, CVPR, ICML and ECCV.

M. Pawan Kumar
http://ai.stanford.edu/~pawan/
Postdoctoral Research Associate
Department of Computer Science
Stanford University
pawan@cs.stanford.edu

Pawan Mudigonda is a post-doctoral researcher at Stanford University. Pawan obtained a Bachelors of Technology degree in Computer Science and Engineering from the International Institute of Information Technology, Hyderabad, India. He completed his PhD studies in 2008 at Oxford Brookes University under the supervision of Prof. Philip Torr and Prof. Andrew Zisserman. His work focuses on combinatorial and convex optimization based solutions for problems in Computer Vision and Machine Learning, and has appeared in several reputed conferences and journals such as ICCV, CVPR, NIPS, ICML and IJCV. Together with his collaborators, he won best paper awards at ICVGIP 2004, Rank Opto-Electronics Symposium 2007, and NIPS 2007.