IMPROVING IMAGE SEGMENTATION BY CHAOTIC NEURODYNAMICS

We propose a novel segmentation algorithm which combines an image segm entation method into small regions with chaotic neurodynamics that has already been clarified to be effective for solving some combinatorial optimization problems. The basic algorithm of an image segmentation i s the variable-shape-block-segmentation (VB) which searches an optimal state of the segmentation by moving the vertices of quadrangular regi ons. However, since the algorithm for moving vertices is based upon st eepest descent dynamics, this segmentation method has a local minimum problem that the algorithm gets stuck at undesirable local minima. In order to treat such a problem of the VB and improve its performance, w e introduce chaotic neurodynamics for optimization. The results of our novel method are compared with those of conventional stochastic dynam ics for escaping from undesirable local minima. As a result, the bette r results are obtained with the chaotic neurodynamical image segmentat ion.