Color image restoration based on framelet regularization

The problem of color image restoration is investigated in this paper. The framelet-based color regularization term is proposed to preserve the edges of restored images and attenuate the false color for human visual system. A fast iterative shrinkage algorithm is applied to tackle the proposed non-smooth restored model by thresholding the framelet coefficients. Numerical examples demonstrate that our fast framelet-based approach is effective to attain the good results with higher PSNR values, lower S-CIELAB errors and less consuming times.