The Research Of Image Enhancement Algorithm Based On The Visual Mechanism Of Retina

In the real world, people acquire a lot of information through the vision, which occupies the vast majority of perceptual information. There is no doubt about the importance of vision for us. Usually the retina samples the light stimulation and process the information to get brightness and color, and then transfer it to the higher visual pathways. The mechanism of information processing in the retina is usually dark adaptation, brightness and color antagonistic etc. The retina cannot have this powerful processing ability without the lateral pathway and the forward pathway which is formed by the wide variety of inner neurons. The forward pathway is responsible for the effective transmission and enhancement of some kind of visual information. In contrast, the lateral pathway link several forward pathways and implement excitatory or inhibitory modulation on the lateral ones.Based on the former findings of physiology of the retina, we proposed several models for image enhancement.1. The retina has a very excellent adaptability on the complex-changing light environment. The triads which are composed of cone cells, horizontal cells and bipolar cells build the physiological foundation of the dark adaptation. This paper proposed a computational model which simulates the response curve of the rod cells and the cone cells, and can modulate the tone of HDRI. The model proves the rationality of the dark adaptation of the retina.2. In the retina, the receptive field of the bipolar cells which have dozens of subspecies firstly tends to be center-peripherals antagonistic. What’s more, the antagonism appears secondly in the ganglion cells. With these mechanisms, the retina can make information in the forward pathways interact with the information in the lateral ones. In this paper, we simulate the dynamic properties of the receptive field of these neurons with variably scaled difference of Gaussian(DOG)model.