Speckle Reduction in Ultrasound Images Using Hybridization of Wavelet-Based Novel Thresholding Approach with Guided Filter

Abstract

AbstractThe research work is aimed to explore and develop some new techniques in the area of speckle reduction. The algorithms are devised and implemented in transform domain for speckle reduction in ultrasound images. The high-quality ultrasound images can improve the accuracy and speed of subsequent image processing tasks and can yield better diagnosis results. In this work, we have proposed a hybrid method to reduce speckle from ultrasound images while maintaining a trade-off between speckle reduction and edge preservation. Extensive work is carried out by taking four different types of images for research which contains synthetic images, simulated images, noise-free ultrasound images, and real ultrasound images. The performance of proposed methods is measured by using different image quality metrics such as PSNR, MSE, COC, SSIM, FSIM, and EPI for all given sets of images except real ultrasound images. The real ultrasound images are compared by using visual comparison and metric SSI because it does not use reference images.KeywordsUltrasound imagesSpeckle noiseWavelet transformGuided filterThreshold method