Optimal Blockwise Shrinkage for Image Denoising

Improving initially denoised using its corresponding noisy observation

This simple script validates the idea of blockwise shrinkage for various image denoising methods. As long as a noisy image is corrupted by additive Gaussian noise, this script method is applicable to improve an initial estimation, namely a denoised image. Demo code shows the improvement based on three MATLAB build-in image denoising methods, i.e. the moving average filter, the median filter, and the 2D wiener filter.