A Image Denoising Algorithm Based on Sparse Dictionary

This paper proposes a sparse denoising algorithm based on noise image training dictionary using K-SVD algorithm to solve the problem that fixed dictionary sparsity is not ideal and the denoising quality is not high enough in image sparse denoising. The main task of K-SVD algorithm is to update the dictionary alternately during iteration process. We compared the image denoising performance of three sparse dictionaries with the images of different types and details. Our results confirm that the image sparse denoising algorithm based on noise image training dictionary has better performance and adaptability than the other two algorithms.