An Improved Non-local Means Algorithm Based on Wave Decomposition

In this paper, we proposed an improved NL-means algorithm based on wave decomposition used for image denoising, the proposed method first employed three-level DWT to the average of the noising image and segmented the noising image to two parts: the textures region and flat region, and then the different matching-window is used for the proposed algorithm. For calculating the weighted kernel more precise, the improved calculating strategy is also presented in this paper. The experimental results prove the proposed algorithm has better performance than the traditional NL-means algorithm, and it can keep the balance between noise suppressing and details keeping.