A novel approach for image denoising based on evolutionary game theory

In this paper, we study the image denoising from the game theoretic perspective and formulate the image denoising problem as an evolutionary game. In the latter, pixels in image are modeled as autonomous players that seek to maximize a payoff function using one of a set of different strategies. The set of strategies of the players are the neighbors. By regarding the non-negative weights of the neighboring pixels as the probabilities of selecting the strategies, the problem of estimating the value of pixels becomes finding the evolutionarily stable strategies for the evolutionary game. Experimental results show that the filtering performance of the proposed approach is very satisfactory and can achieve better performance than the median filter in terms of PSNR, MSSIM and visual quality.