Research On Color Image Enhancement Techniques

Image enhancement is an important part in image processing, its role is to improve the image's quality to meeting specific applications. For the sake of efficient storage and transmission, increasing images are being represented in the compressed format. Thus, more and more attentions are shifted to compressed domain image enhancement.This paper presents several classical color image enhancement algorithms in DCT compressed domain as well as their advantages and disadvantages, especially the recently proposed CES algorithm by Mitra is discussed in details. We found that CES algorithm produces blocking artifacts in the boundaries of image blocks, which seriously affect the image's quality. This paper analyzes the reasons why CES produces blocking artifacts, and proposes an improved algorithm- ICES algorithm, which employs anisotropic diffusion filter to suppress the blocking ar-tifacts. Simulation results show that, without increasing the computations, ICES algorithm can not only suppress blocking artifacts effectively, but also maintain comparable enhancement performance. Otherwise, ICES algorithm also possesses the advantages as the CES algorithm, such as preserving color constancy, and directly extending to JPEG, MPEG and H.26x compressed standards, etc., with overall higher performance than CES algorithm.