Image Segmentation by Co-existing Strange Attractor Networks

Based on the framework of concurrent synchronization of dynamical systems, the Newton-Leipnik (NL) chaotic oscillator network with co-existing strange attractors, are proposed to solve image segmentation. The N-L oscillators whose underlying visual stimulative atoms belong to the same visual group synchronize more quickly compared with F-N neural oscillators, and the segmentation obtained by N-L oscillators seems more subtle using the same number of classes. The simulation has shown that the N-L co-existing strange attractor network achieves promising results on image segmentation.