Speckle filtering of ultrasound B-Scan Images - a comparative study between spatial and diffusion filters

Abstract

Speckle noise is the inherent property of ultrasound B-Scan images which has been filtered using well-established speckle reduction techniques. In this work, six spatial filters namely Frost, Median, Lee, Kuan, Wiener, and Homomorphic filters, and two diffusion filters viz., Speckle Reduction Anisotropic Diffusion (SRAD) filter, and Anisotropic Diffusion (AD) filter have been attempted over 200 different digital ultrasound B-scan images of kidney, abdomen, liver and choroids. A comparative study has been made on these filters in preserving the edges of the images with effective denoising by calculating fourteen established performance metrics along with the execution time in order to determine the effective and optimum despeckling algorithm for real time implementation. To do this, a cumulative speckle reduction (CSR) algorithm has been developed using MATLAB 7.1, which performs all despeckle filtering functions as well as performance metrics calculation in a single iteration. This study reveals that most of the despeckle filters performed well and gave optimum performance, but SRAD is the outperformed filtering technique for B-scan ultrasound image as far as the performance metrics, execution time and visual inspection are concerned.