Adaptive retinex back-light compensation algorithm using skewness information of image

초록

This paper presents an adaptive Retinex algorithm. In order to solve the typical problems of the Retinex algorithm, such as expensive computational cost, halo artifact, and color distortion, a function of skewness that represents a statistical distribution of pixels is defined to compensate for contrast and color distortion. The experimental results show that the proposed algorithm leads to subjectively better performance than the typical Retinex algorithm, and that proposed algorithm also has the capability of reducing the computational cost by approximately 40%. © 2013 ACM.