We present a full-wave inversion approach with total variation regularization for elastography. The proposed method is based on the minimization of an error in constitutive equations functional augmented with a least squares data misfit term referred to as MECE for “modified error in constitutive equations.” The main theme of this paper is to demonstrate several key strengths of the proposed method on experimental data. In addition, some illustrative examples are provided where the proposed method is compared with a common shear wave elastography (SWE) approach. To this end, ultrasonically tracked displacement data from an acoustic radiation force (ARF) pulse are used in different phantoms including phantom with layered inclusion and triangle inclusion. The results indicate that the MECE approach can produce accurate shear modulus reconstructions in comparison with SWE, especially around the sharp edges in the layered and triangle inclusions. We compare shear modulus reconstruction using MECE and SWE with original inclusion shapes using two-dimensional normalized zero mean cross correlation, edge preservation index and dice coefficient similarity index. [Work supported by NIH Grant R01 CA174723.]
Non-Local Based Denoising Framework for In Vivo Contrast-Free Ultrasound Microvessel Imaging
Vascular networks can provide invaluable information about tumor angiogenesis. Ultrafast Doppler imaging enables ultrasound to image microvessels by applying tissue clutter filtering methods on the spatio-temporal data obtained from plane-wave imaging. However, the resultant vessel images suffer from background noise that degrades image quality and restricts vessel visibilities. In this paper, we addressed microvessel visualization and the associated noise problem in the power Doppler images with the goal of achieving enhanced vessel-background separation. We proposed a combination of patch-based non-local mean filtering and top-hat morphological filtering to improve vessel outline and background noise suppression. We tested the proposed method on a flow phantom, as well as in vivo breast lesions, thyroid nodules, and pathologic liver from human subjects. The proposed non-local-based framework provided a remarkable gain of more than 15 dB, on average, in terms of contrast-to-noise and signal-to-noise ratios. In addition to improving visualization of microvessels, the proposed method provided high quality images suitable for microvessel morphology quantification that may be used for diagnostic applications.