在超声成像中利用 Eikonal 方程结合有限元法进行传播时间估计的计算机模拟研究
An in-silico study of travel time estimation using Eikonal equation with finite element method in ultrasound imaging
AI总结:
本研究提出在超声成像中用有限元法求解 eikonal 方程以估计非均匀介质中的超声传播时间,模拟实验表明其性能与 fast marching 方法相当,且具备高阶扩展和 GPU 加速潜力。
摘要:
估计超声在非均匀介质中的传播时间对于高质量成像至关重要,因为借助对声速分布的准确估计,通常被视为图像退化原因之一的相位像差便可通过多种方法得到校正。为了估计超声在非均匀介质中的传播时间,利用 eikonal 方程解的方法颇具吸引力,因为它可以方便地集成到传统的延迟求和波束形成器中。在本文中,我们首次提出采用有限元法求解超声成像中的 eikonal 方程,该方法易于扩展到高阶数值格式,并有可能在 GPU 上更快运行。通过一系列计算机模拟实验,我们所得结果与 fast marching 方法的结果相当。
英文摘要:
Estimating the travel time of ultrasound in an inhomogeneous medium is crucial for high-quality imaging, as with an accurate estimate of the distribution of speed of sound, phase aberration, which is normally viewed as one of the reasons for image degradation, can be corrected by multiple methods. In order to estimate the ultrasound travel time in an inhomogeneous medium, the approach using solutions to the eikonal equation is attractive, since it can be easily integrated into a conventional delay-and-sum beamformer. In our manuscript, we first propose employing the finite element method to solve the eikonal equation in ultrasound imaging, which is easy to extend to a high-order numerical scheme and potentially runs faster on GPU. Through a series of in-silicon experiments, our results are comparable to those of fast marching methods.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。