A New Method of an Improved PCNN Model for Image Segmentation Based on Fuzzy Entropy

A new improved PCNN model is proposed to overcome the problem existing in the image segmentation of PCNN.It doesn't only simplify the acceptant part of the traditional PCNN and improves on the pontes of PCNN and changes the threshold attenuation mode of PCNN,but also utilizes the maximum fuzzy entropy as the determinant rule of the best segmentation iterations.Therefore,the improved PCNN can implement the segmentation results accurately and automatically.Simulation experiment on various kinds of images indicates that the proposed method can confirm the circulatory iterations and choose the best threshold automatically.Compared to the maximum Shannon entropy,the proposed method has higher convergence speed and segmentation accuracy,and also achieves a better segmented effect.