Homomorphic Filtering and MAD Filtering Based Speckle Removal in Ultrasound Images

Abstract

The economy of medical ultrasound images comes at the expense of image quality that is usually affected by speckle noise. In this paper, a method for 2D ultrasound medical image despeckling is presented, which combines homomorphic filtering and MAD (Median Anisotropic Diffusion) filtering that the improved Anisotropic Diffusion filter. First, the preprocessed image after dilution of speckle-noise is obtained by homomorphic transformation of the ultrasound image with Gaussian high-pass filtering and median filtering. Then the speckle is removed with the MAD filtering, and finally, the contrast is enhanced. Experimental results demonstrate that the algorithm can effectively reduce the speckle noise and maintain high image quality. The algorithm in this paper helps remove speckle in 2D static ultrasound images, which is potentially valuable for improving the quality of medical ultrasound images.