BM3D Image Denoising using SSIM Optimized Wiener Filter

Image denoising is considered as a salient pre-processing in sophisticated imaging applications. Over decades, numerous studies have been conducted in denoising. Recently proposed Block Matching and 3D (BM3D) Filtering added a new dimension to the study of denoising. BM3D is the current state-of-the-art of denoising and is capable of achieving better denoising as compared to any other existing method. However, the performance is not yet on the bound for image denoising. Therefore, there is scope to improve BM3D to achieve high quality denoising. In this thesis, to improve BM3D, we first attempted to improve Wiener filter (the core of BM3D) by maximizing the Structural Similarity (SSIM) between the true and the estimated image, instead of minimizing the Mean Square Error (MSE) between them. Moreover, for the DC-Only BM3D profile, we introduced a 3D zigzag thresholding. Experimental results demonstrate that regardless of the type of the image, our proposed method achieves better denoising than that of BM3D