October 19, 2012
Koenderink Prize for Vladimir Kolmogorov
IST Austria Professor honored for paper with long-term impact and fundamental contributions to computer vision
Vladimir Kolmogorov, Professor at the Institute of Science and Technology (IST) Austria, received the Koenderink Prize at this year’s European Conference on Computer Vision (ECCV) in Florence (October 7-13) for fundamental contributions to the field of computer vision. The Koenderink Prize is awarded each year at the ECCV, one of the most prestigious conferences in the field of computer vision worldwide, for papers published at ECCV ten years ago which have withstood the test of time.
Kolmogorov received the Koenderink prize for his paper “What Energy Functions can be Minimized via Graph Cuts?”, written during his doctoral studies with Ramin Zabih at Cornell University in 2002. Graph cuts is an optimization technique that has played a major role in computer vision in the last decade. The work of Vladimir Kolmogorov on graph cuts is well-known in the community and extensively cited. This particular paper brought the concept of submodularity to computer vision; it characterizes conditions under which graph cuts can be applied. Graph cuts are used to efficiently solve various problems in computer vision, in particular those that can be formulated in terms of energy minimization. An interactive image segmentation algorithm based on graph cuts, called “Grabcut”, has been incorporated in Microsoft Office 2010. Grabcut uses graph cuts to separate an image into its foreground and background with only very limited input by the user, restricted to drawing a rectangle around the object of interest.