Adaptation of Telegraph Diffusion Equation for Noise Reduction on Images

Abstract

The Telegraph Diffusion Equation (TDE) used in some noise reduction processes in an image includes two main parameters: the damping coefficient and the relaxation time. Classically, the first is determined globally for a given input image, while the second one is set constant. In this paper, we propose to determine the values of these parameters according to the information and the image local structure. We then get an adaptive diffusion equation that permits to better control the degree of smoothness and preserve fine structures and image contours avoiding speckles phenomena and staircase. The acquired results show that the proposed method improves the quality of images that have a weak signal-to-noise ratio, comparatively to the methods based on the TDE whose parameters are not adaptive.

Hybrid model of diffusion based on the Jeffreys‐type equation for noise reduction on images

“…11 a and b show the PSNR of the images of Lena and bridge on which were applied different standard deviation of the noise (10, 20, 30) in function different values of . This technique for determining is commonly used in [9, 10]. The selected value .…”

“…The main limitation of their method is the calculation time of the bilateral filter and the choice of the damping coefficient in the TDE which does not depend on the image features. Ghislain et al [10] proposed the adaptation of TDE (A‐TDE). They showed that the damping coefficient and the relaxation time depend on the image features.…”

Image Similarity Test Using Eigenface Calculation

“…In the eigenface method, decoding is done by calculating the eigenvector and then represented in a large matrix [2]. This matrix has a feature value that serves to assess how close to a comparable object.…”