Research Article
Image Denoising via Asymptotic Nonlocal Filtering
(1) Initialization: . | (2) Decomposition: Decompose into and () by wavelet decomposition. | (3) De-noising: | (A) Reconstruct the low and high-frequency image , respectively from and by inverse wavelet | transform. And let ; | (B) Calculate similarity weights by (12) on the low sub-image ; | (C) Estimate by (11) on the sub-image ; | (D) Replace by ; | (E) Repeat the steps A–D at each layer until the layer and obtain restored image . |
|