Khaled Alyemni, summer 2016, "Iterative Adaptive Non-Local Means for Image Denoising", Computer Science Department, Faculty of Science, University of Western Ontario, Canada

M.Sc. Thesis Abstract

The Non-Local Means (NLM) method of denoising has received considerable attention in the image processing community due to its performance, despite its simplicity. In this research, a new version of NLM algorithm is present that is based on NLM-SAP. The proposed algorithm is an iterative version on NLM-SAP algorithm. NLM-SAP uses different shapes of patches rather than square shape patches in denoising images. Moreover, NLM-SAP uses a faster approach than original NLM which makes image denoising less computationally intensive. The proposed approach in this research is an iterative adaptive approach that applies the NLM-SAP algorithm to the noisy image and changes the parameters in each iteration based the noise level. In more details, the algorithm decreases the values of the parameters since the noise is decreasing in each iteration. The iterative approach stops when there is no improvement in the image. The result shows that the proposed approach outperforms NLM-SAP on different images.