Multi-Color Channels Based Group Sparse Model for Image Restoration

The group sparse representation (GSR) model combines local sparsity and nonlocal similarity in image processing, achieves excellent results. However, the traditional GSR all subsequent improved models convert RGB space of to YCbCr space, only extract Y (luminance) channel change color a gray for processing. As result, processing process cannot be loyal each channel, so repair effect is not ideal. A new based on multi-color channels proposed this paper. processes R, G B simultaneously when images rather than single then combining results different channels. multi-color-channels-based compared with state-of-the-art methods. experimental contrast show that an effective method can obtain good terms objective quantitative metrics subjective visual effects.