The application of nonlocal total variation in image denoising for mobile transmission

Image transmission is one of the key techniques in image mobile communication. However, it is generally corrupted by noise in wireless channel, which will decrease the visual quality and affect the sub-sequential applications, such as pattern recognition, classification and so on. Total variation is widely used in the problems of image denoising, due to its advantage in preserving texture in image. In this paper, a novel minimization framework is presented where the objective function includes an usual l2 data-fidelity term and two types of total variation regularizer. According to the theory analysis, the novel objective function can preserve the local geometric structure in restored image. Furthermore, we proposes to solve the novel framework with majorization- minimization and compares this novel algorithm with some current restoration method. The numerical experiments show the efficiency and effectiveness of the proposed algorithm.