Segmentation Averaging with Applications in Shape Matching and Perceptual Organization


November 24, 2008


Hongzhi Wang


Stevens Institute of Technology


We use segmentations to match images by shape. To address the unreliability of segmentations, we give a closed form approximation to an average over all segmentations. Our technique has many extensions, yielding new algorithms for tracking, object detection, segmentation, and edge preserving smoothing. For segmentation, instead of a maximum a posteriori approach, we compute the “central” segmentation minimizing the average distance to all segmentations of an image. Our methods for segmentation and object detection perform competitively, and we also show promising results in tracking and edge-preserving smoothing.


Hongzhi Wang

Hongzhi Wang is a 5th PhD student majored in computer science at Stevens Institute of Technology. His research interestes include perceptual organization, image segmentation, object recognition and Bayesian methods. He obtained his Bachelor and Master degree in computer science from University of Science and Technology Beijing in 2000 and 2003.