Image segmentation based on PCNN model combined with automatic wave and synaptic integration

A PCNN model combined with synaptic integration and automatic wave is presented in this paper. The fired neurons and unfired neurons in neighborhood are taken as excitatory and inhibitory synapses respectively, and the result of synaptic integration serves as the PCNN linking input; the firing map of the image spreads in decaying automatic wave, then the segmentation result is obtained when the map turn to be stable. The experimental results demonstrate the proposed model perform well in edge areas and restrains the over segmentation phenomenon, the shape measure and the contrast measure are improved at the same time.