James-Stein Type Center Pixel Weights for Non-Local Means Image Denoising

摘要

Non-Local Means (NLM) and variants have been proven to be effective androbust in many image denoising tasks. In this letter, we study the parameterselection problem of center pixel weights (CPW) in NLM. Our key contributionsare: 1) we give a novel formulation of the CPW problem from the statisticalshrinkage perspective; 2) we introduce the James-Stein type CPWs for NLM; and3) we propose a new adaptive CPW that is locally tuned for each image pixel.Our experimental results showed that compared to existing CPW solutions, thenew proposed CPWs are more robust and effective under various noise levels. Inparticular, the NLM with the James-Stein type CPWs attain higher means withsmaller variances in terms of the peak signal and noise ratio, implying theyimprove the NLM robustness and make it less sensitive to parameter selection.