Wavelet Based Image Denoising Using Self Organizing Migration Algorithm
Anupriya Anupriya, Akash Tayal · CiiT international journal of digital image processing · 2012
Wavelet shrinkage denoising is a popular method fornoise removal in images. The parameter selection for this technique is dependent on the statistics of the original image and noise. Usually these parameters are chosen by trial and error. The first technique introduced for determination of threshold for wavelet shrinkage was the universal threshold proposed by Donoho. However, this technique does not give satisfactory results in all cases. In this paper, we propose the use of Self-Organizing Migration Algorithm (SOMA) for finding the optimal set of parameters (type of wavelet, level of decomposition and threshold value) for wavelet shrinkage denoising. SOMA generates good quality solution, for various kinds of noises (gaussian, speckle and salt & pepper noise), in a reasonable amount of time. Evaluation of denoised image is done using parameters such as PSNR, entropy and SSIM. It is shown that proposed method gives high degreeof noise removal while preserving the edges and other details in the image.