Minimal-complexity segmentation with a polygonal snake adapted to different optical noise models

Polygonal active contours (snakes) have been used with success for target s egmentation and tracking. We propose to adapt a technique based on the mini mum description length principle to estimate the complexity (proportional t o the number of nodes) of the polygon used for the segmentation. We demonst rate that, provided that an up-and-down multiresolution strategy is impleme nted, it is possible to estimate efficiently this number of nodes without a priori knowledge and with a fast algorithm, leading to a segmentation crit erion without free parameters. We also show that, for polygonal-shaped obje cts, this new technique leads to better results than using a simple regular ization strategy based on the smoothness of the contour. (C) 2001 Optical S ociety of America.