Threshold Estimation Based on Perona-Malik Model

The selectivity of diffusion mechanism and parameters is the key in the field of image nonlinear diffusion filtering based on Perona-Malik model. The threshold estimation of the diffusion function depends mainly on the experience in the existing methods. In this paper, the relationship between the image features, the noise variance and the threshold are analyzed, and a new threshold estimation method is proposed, the test results show it is effective.