A Retina-Filtering Model and its Application in Image Restoration

It is one of the attractive and challenging topics in the field of computer vision to reveal the mechanism of the biological visual system and gain useful insights from it. According to the experimental results of physiology, this paper analyzes the role of the image transmission components in mammalian retina for improving imaging clarity. Then, based on the theory of Bayesian estimation, a kind of retina-filtering model is proposed, of which the parameters were determined by Green's solution of fractional diffusion equation and the optical characteristics of eye. Finally, experiment results in image restoration show the effectiveness and feasibility of the proposed method.