Wavelet based Denoising of Medical Images Using Sub-band Adaptive Thresholding through Genetic Algorithm

Abstract

Abstract Generally medical images have poor contrast along with serious types of noises. The suppression of noise in medical images corrupted by Gaussian white noise is a major issue in diverse image processing and computer vision problems. Image denoising using discrete wavelet transform is well established domain in image processing because it can separate the noisy signal from the image signal. This paper proposed a denoising method of medical images through thresholding and optimization using a stochastic and randomized technique of Genetic Algorithm (GA). The noisy image is partitioned into fixed sized blocks and then transforms it into wavelet domain. Some important parameters in the 2-D discrete wavelet transform such as the decomposition level and the threshold value are searched and optimized in a wide range in the proposed technique. The Bayesian shrinkage method has been selected for thresholding based of its sub band dependency property. Proposed algorithm has been validated through ultrasound image corrupted by a variety of noise densities through Gaussian noise in terms of peak signal to noise ratio and visual effects. Simulation results show that the proposed method outperforms the existing denoising methods.