Subsquarewise Threshold Based Image Denoising

A new subsquarewise threshold which value is related to the coefficients in the subsquare is proposed in this paper. By modeling the bandelet coefficients in each subsquare as a generalized Gaussian distribution (GGD) model, we propose a denoising algorithm based on the second generation orthogonal bandelets and subsquarewise threshold. It can guarantee the optimal approximation to edge by using the second generation bandelet which can adaptively capture geometry of regular image along its edges, and can also guarantee the effective denoising by using subsquarewise threshold. Experimental results show that the proposed algorithm performs well in terms of both the visual effect and the objective evaluation criteria.