Wavelet based Image Denoising with Locally Adaptive Window and Non-Local Means Filter Vijayaraghavan V.*, Dr. Karthikeyan M.** *Research Scholar, Information and Communication Engineering, Anna University, Chennai, India **Principal, Tamilnadu College of Engineering, Coimbatore, India Online published on 23 March, 2017. Abstract Denoising is the fundamental step in image preprocessing. It is necessary because image is often contaminated by noise during acquisition and transmission. Denoising process removes noise by retaining all other characteristics of image. In the proposed method Discrete Wavelet Transform (DWT) is used that has the ability to capture energy of the signal into few energy transform value. The threshold value is computed by estimating the noise and signal variance from the transformed image. Then thresholding the transformed image is done by using a Locally Adaptive Window Maximum Likelihood (LAWML) shrinkage function, which undergoes Non-Local Means (NLM) filter to improve the denoising performance. Evaluation is carried out in terms of PSNR. Experimental results over a range of noise power levels indicate that the proposed method is better than other denoising methods. Top Keywords Image denoising, locally adaptive window, non-local means, PSNR, wavelet shrinkage, wavelet transform. Top |