Skip to Main Content
Image enhancement is used to improve quality of images for further analysis. In this paper, we proposed a color image enhancement method with the use of sparse representation of orthogonal transform based on the characteristics of discrete cosine transform (DCT) coefficients: fewer nonzero coefficients exist in smooth domains, the amplitudes of DCT change slightly in texture contents but significantly in edge. Nonlinearly reduces the impact of textures with the use of the threshold on DCT coefficients, and then enhances the texture and edge regions by applying the a-power computation on nonzero coefficients in medium and high frequency domains. To remove blocky effects, Kaiser window function is used on each processed image block. Compared with traditional method, this image enhancement algorithm achieves better visual effects, while the noise in texture region is restrained at the same time.