De-Speckling by Hybridization of Wavelet Transform and Bilateral Filter in Ultrasound Images

Abstract

In this paper we introduce a de-noising technique with hybridization of wavelet transform and bilateral filter. In digital image processing, image is corrupted by various types of noise. But medical images are usually corrupted by speckle noise. So to remove noise, image de-speckling is very essential exercise of diagnose. To speckle noise reduction, multiplicative noise is converted into additive noise. Various filters are implemented for the reduction of speckle noise from ultrasound images. The proposed method uses wavelet transform and bilateral filter to reduce speckle noise from ultrasound images with homomorphic approach. We compare our results with mean filter, median filter, Gaussian filter, bilateral filter, wiener filter, Lee filter, Kuan filter and SRAD filters. So, we use bilateral filter using homomorphic approach with wavelet thresholding. The use of Weiner filter and soft thresholding provides better results to reduce speckle noise and gives better image quality with fine details. Another method is developed for mixed noise reduction from ultrasound images. We compare the results of proposed methods with the existing approaches in terms of PSNR and MSE. The hybridization of wavelet transforms and bilateral filter has been shown work well than other smoothing filters and mixed noise reduction methods.