Based On The Dynamic Characteristics O F The Reception Field Research Of Image Enhancement Algorithm

Image enhancement can improve image quality, which is required to highlight theimage information we need according to specific requirements and weaken or removethe image information we don’t focus on. The human visual information processingmechanism can effectively enhance image. Early studies suggest that receptive field ofretinal ganglion cells and lateral geniculate cells present concentric circles structure, andits main function is to test the brightness contrast of the image and extract the edgefeature of the image. But, there is fatal flaw when using it for image enhancement: thebrightness gradient information of the image (Low-frequency components) would befiltered.In this paper, two RF-dynamic based adaptive image enhancement methods areproposed. The first method is based on the three Gaussian function model of receptivefield, which uses the dynamic property of receptive field to select the most appropriatethree Gaussian function model to deal with the local area of an image according to thedifferent local contrast among the image. Then, we can get better image enhancementresults. Compared with Histogram equalization, our method can effectively improve theedge feature and the local details of the image.The second method is based on the subunit model of receptive field of retinalganglion cells and lateral geniculate cells. The core mechanism of subunit model is thatthe mutual inhibition of all the subunits of receptive field leads to disinhibition.According to the dynamic property of receptive field, this method can indirectlyregulate the mutual inhibition strength of all the subunits through dynamically changingthe radius size of the inhibition area of receptive field. Then, it can realize adaptiveimage enhancement.