Image denoising method based on PM model with transforming edge stopping function
Jimin Yu, Rumeng Zhai, Jiayong Yie · 2017 IEEE 2nd Information Technology, Networking, Electronic and Automation Control Conference (ITNEC) · 2017
Owing to the anisotropic Perona-Malik(PM) diffusion model is often used for image denoising and enhancement, and in this model the diffusion rate is controlled by the edge stopping function, the significant shortcoming of PM model is the failure in preservation of sharp edges and textures. In this paper, we mainly analyze the effects of different edge stopping functions with regard to edge protection. By selecting an appropriate edge stopping function, the drawback of the PM model mentioned above can be overcome. Therefore, in this paper, the characteristics of several different edge stopping functions are discussed, and finally some guidance suggestions are given for how to choice appropriate edge stopping function and the corresponding gradient threshold K.