Single Image Super Resolution via Adaptive Group-Based Sparse Domain Selection

Abstract

sparse representation has been used as a powerful statistical image modelling technique in single image super resolution (SISR). Although this prior efficiently is utilized to describe the local smoothness but ignoring the correlation between the sparse representation coefficients of similar patches can lead inaccurate spare coding coefficients. In this paper, we propose the method that enforce the local smoothness and nonlocal self-similarity by sparse representation in a unified framework, called adaptive group-based sparse domain selection (A-GSDS). N onlocal patches with similar structures are leveraged and stacked into a matrix as the basic unit of sparse representation called group. These groups are converted into a column vector, each column selects the best fitted PCAA sub dictionary learned from the training data. After applying the sparse coding process to each column in the group domain, sparse vectors are obtained by orthogonal sub dictionaries which can be easily estimated. To further improve the performance of the group-based sparse representation, we use nonlocal means regularization term. Extensive experimental results validate the effectiveness of the proposed method comparing with the state-of-the-art algorithms.