Image Denoising via Group Sparse Eigenvectors of Graph Laplacian

In this paper, a group sparse model using Eigenvectors of the Graph Laplacian (EGL) is proposed for image denoising. Unlike the heuristic setting for each image and for each noise deviation in the traditional denoising method via the EGL, in our group-sparse-based method, the used eigenvectors are adaptively selected with the error control. Sequentially, a modified group orthogonal matching pursuit algorithm is developed to efficiently solve the optimal problem in this group sparse model. The experiments show that our method can achieve a better performance than some well-developed denoising methods, especially in the noise of large deviations and in the SSIM measure.