Video Image Restoration Based on Sparse Representation

Abstract

The sparse representation, as a powerful model, has been successfully used in image processing. Considering that the sparse representation can be used in static image restoration, we propose a video image restoration method on sparse representation. In our algorithm we use a ternary function to express a video image. Adaptively selecting the dictionary by using the Principal Component Analysis (PCA) strategy and getting the sparse coding coefficients by the continuous iteration, finally we get the restored video image. The restored results by our method are compared with those by the other three methods. They are the wiener filter, the blind deconvolution, and the learning model method of natural image patches [7]. In order to better assess the quality of the video image restoration we use both the subjective evaluation and the objective evaluation. Extensive experiments show that our proposed method is effective and superior to the other methods.