Realistic image rendition algorithm based on Sigmoid-iCAM color appearance model

In order to simulate the imaging mechanism of human vision system, reproduce the same scene perception of human eyes and provide machine vision the ideal images which are in accordance with biological visual characteristics, color image reproduction technology is studied from the colorimetry. Aiming at the shortcomings of image color appearance model fuzzy and low contrast, the visual local adaptation mechanism based on the Sigmoid function of artificial neural network is introduced, then, a realistic image rendition algorithm based on Sigmoid-iCAM color appearance model is proposed. Compared with the realistic image rendition algorithms based on CIECAM02 and iCAM models, this algorithm can improve the image contrast and local details effectively, and can achieve a good color restoration for color images with color distortion, and can well realize the visual system of color constancy.