Edge detection with the Improved Pulse Coupled Neural Network using statistical features

Recently, Pulse Coupled Neural Network is effectively used for the image processing such as image thresholding, image segmentation and edge detection. We propose an improved PCNN model by reclaiming the input of it. In the proposed method, we use the statistical features as input of the PCNN. This method was suggested by modifying the parameters using the neighboring information including, mean, variance, kurtosis and skewness. The experimental results show that the feature based method gives the improved performance metric and enhanced the quality of the edge detection rather than the pixel-based one. Our findings on a benchmark dataset demonstrate that the proposed method outperforms than other edge detection techniques particularly regarding the basic PCNN.