Application of Demons algorithm in ultrasound elastography using B-mode ultrasound images

Abstract

UltraSound Elastography (USE) has been widely used to obtain mechanical properties of tissues. Radio frequency (RF) data is usually used in USE to estimate the displacement. However, RF data is not available in all ultrasound imaging devices. B-mode images which are basically the envelope of the RF data are the most well-known output of the common ultrasound imaging devices. In B-mode images, the phase information of RF data is lost. Consequently, USE can be more challenging and the strain image quality would be degraded. The aim of this paper is to employ Demons algorithm, which is a powerful non-rigid image registration algorithm, to estimate displacement using B-mode images. In USE, the post-compression image may have large deformations in axial direction which deteriorates the Demons algorithm performance. In order to compensate the large deformations, an optimization algorithm is proposed to find and compensate the mean value axial deformation. Experimental and numerical phantoms are used to verify the algorithm performance in normal and severe situations. The results are compared with the common normalized cross correlation (NCC) algorithm. The results confirm that Demons algorithm is an appropriate algorithm for USE for B-mode images considering the fact that phase information are not available.