A New Image Segmentation Algorithm Based on PCNN and Maximal Correlative Criterion

Pulse Coupled Neural Network (PCNN) is a new generation of artificial neural networks, which has biological background, embodies excellent performance in image segmentation. However, the problem of parameter estimation and threshold iteration in PCNN model has not been resolved yet. This paper combined 1-dimensional Maximal Correlative Criterion with 2-dimensional Maximal Correlative Criterion to estimate neuron parameters, achieved the automation of image segmentation and reduced the complexity of computing. Simulation results showed that the algorithm has prominent improvement in image segmentation effect and computing complexity and has general applicability compared to relevant literatures.