In order to removes the ring effect in traditional image denoising algorithms using wavelet thresholding, the paper analyzes the wavelet coefficients of noise images, uses second-order central moment of HH1 sub-bands as the noise variance and computes threshold values; and then performs wavelet thresholding denoising on each image block. At last, the paper weights these denoised wavelet coefficients using Kaiser Function. Compared with traditional algorithms, because the algorithm in the paper makes use of the characteristics of wavelet coefficients and the Human Vision System Theory, it preserves the edge and texture information and removes the ring effect. Moreover, the algorithm uses Kaiser Function to overcome blocking effects and improves the visual effects of the denoised image.

Image Wavelet Denoising based on Human Vision System Theory

Mixed noise composed by Gaussian noise and Salt-Pepper noise is an ever-present noise model in the image. This paper proposes an adaptive image mixed removal algorithm based on measure of medium truth degree(MMTD). According to the feature and density of noise, it adaptively alters the detection window size. Then it defines the predicates and establishes the relation between gray level and truth interval of predicates. Finally, uses the distance ratio function to measure the similarity degree between the considered pixel and the normal pixel in the detection window and to remove the noise pixel. By sample simulation and classic PSNR evaluation, it shows the adaptive image mixed noise removal algorithm (AdpMMTD) brings about a good performance in removing mixed noise and preserving fine details.

An adaptive image mixed noise removal algorithm based on MMTD

The medium mathematics system is another mathematical tool which deals with fuzzy and uncertain problem. According to the analysis of the features of the image mixed noise, this paper introduces a new image mixed noise removal algorithm based on measuring of medium truth scale. It uses the distance ratio function to detect the noise pixel and to restore the image. The experimental results demonstrate that the new image mixed noise removal algorithm can do better in smoothing mixed noise and preserving details than the classical ones do in subjective aspect and objective aspect, which will lead to its practicable and effective applications in mixed noise removal and image restoration.

Long Hong

Papers

A new image mixed noise removal algorithm based on measuring of medium truth scale