Optimization of Breast Displacements Estimation in Ultrasound Elastography

Abstract

Static ultrasonic breast elastography is a medical imaging technique that produces information related to the elastic properties of breast tissue. However, the image quality in breast elastography is considerably altered by the diffusion of speckle noise in the breast texture leading to disrupt the medical diagnosis. In this perspective, the implementation of a technique for optimizing the tissue displacements estimation is crucial. Therefore, in this paper, we proposed a new method based on the monogenic model, in order to improve the Classical Monogenic Signal (CMS) technique by improving the filtering process, in this context, a filtering step has been applied to the B-mode images and then the filtered B-mode images were integrated into a monogenic model to estimate displacements. Our study was carried out on two models of phantom elasticity sold by the CIRS Company. In vivo 20 B-mode images were also performed corresponding to 20 patients including malignant breast tumors. An accurate and excellent breast tissue displacements estimation image was noted in the results of the proposed technique. Also we noticed the good quality and preservation of the edges tissue. The proposed method provides better values for standard deviation (SD) calculation, greater contrast to noise ratio (CNR), higher peak signal-to-noise ratio (PSNR), excellent structural similarity (SSIM) and much more Fast than CMS and B-Spline (BS). The results of the proposed model are promising and efficient, allowing a rapid estimate. The proposed strategy makes it possible to improve the diagnosis of breast pathologies and helps to establish a more precise and advanced clinical evaluation of the diseases affecting the breast.