A Modified Wiener Filtering Method for DT-CWT Domain Image Denoising

A novel scheme for reducing additive Gaussian noise from images in the dual tree complex wavelet transform (DTCWT) domain is introduced in this paper. The doubly Wiener filtering algorithm is investigated and a simple one-step Wiener filtering method in DT-CWT domain is derived by choosing optimal parameter to minimize the local expected square error. Comparing with the current best denoising techniques, the proposed method outputs comparable results (visually and PSNRs) but requires far lower computations. The experimental results demonstrate the effectiveness of the proposed scheme.