Towards a comprehensive evaluation of ultrasound speckle reduction

Abstract

Over the last three decades, several despeckling filters have been developed to reduce the speckle noise inherently present in ultrasound images without losing the diagnostic information. In this paper, a new intensity and feature preservation evaluation metric for full speckle reduction evaluation is proposed based contrast and feature similarities. A comparison of the despeckling methods is done, using quality metrics and visual interpretation of images profiles to evaluate their performance and show the benefits each one can contribute to noise reduction and feature preservation. To test the methods, noise-free images and simulated B-mode ultrasound images are used. This way, the despeckling techniques can be compared using numeric metrics, taking the noise-free image as a reference. In this study, a total of seventeen different speckle reduction algorithms have been documented based on adaptive filtering, diffusion filtering and wavelet filtering, with sixteen qualitative metrics estimation.