An Investigation of Denoising Parameters Choice in two Perona-Malik Models

The paper addresses the problem of no-reference parameter choice for image denoising by Perona-Malik image diffusion algorithm using two models. The idea of the approach is to analyze the difference image between noisy input image and the outcome of the denoising algorithm for the presence of structured data from the input image. The analysis consists of the calculation of the mutual information — a value that shows the ratio between the structured data and the noise. We apply the proposed method to photographic images, vector graphics images and to retinal images with modeled Gaussian noise with different parameters.

[1]  Santiago Aja-Fernández,et al.  On the estimation of the coefficient of variation for anisotropic diffusion speckle filtering , 2006, IEEE Transactions on Image Processing.

[2]  Max A. Viergever,et al.  Ridge-based vessel segmentation in color images of the retina , 2004, IEEE Transactions on Medical Imaging.

[3]  Jitendra Malik,et al.  Scale-Space and Edge Detection Using Anisotropic Diffusion , 1990, IEEE Trans. Pattern Anal. Mach. Intell..

[4]  Eero P. Simoncelli,et al.  Image quality assessment: from error visibility to structural similarity , 2004, IEEE Transactions on Image Processing.

[5]  Maria Petrou,et al.  On the choice of the parameters for anisotropic diffusion in image processing , 2013, Pattern Recognit..

[6]  Andrey S. Krylov,et al.  Choice of the Parameter for BM3D Denoising Algorithm Using No- Reference Metric , 2018, 2018 7th European Workshop on Visual Information Processing (EUVIP).

[7]  J. Koenderink.  The structure of images , 2004, Biological Cybernetics.

[8]  Karl Krissian,et al.  Noise-Driven Anisotropic Diffusion Filtering of MRI , 2009, IEEE Transactions on Image Processing.

[9]  Nikolay N. Ponomarenko,et al.  Image database TID2013: Peculiarities, results and perspectives , 2015, Signal Process. Image Commun..