INVERSE FINITE ELEMENT BASED ANISOTROPIC MATERIAL PROPERTY RECONSTRUCTION METHO

该子项目是利用该技术的众多研究子项目之一 资源由 NIH/NCRR 资助的中心拨款提供。子项目及 研究者 (PI) 可能已从 NIH 的另一个来源获得主要资金， 因此可以在其他 CRISP 条目中表示。列出的机构是 对于中心来说，它不一定是研究者的机构。 该项目的目标是通过剪切波位移测量获得软组织的各向异性弹性特性。这将通过首先使用有限元方法模拟正演问题（即弹性材料中的剪切波传播）来实现。其次，将根据弹性成像实验的测量位移和正问题的预测位移建立逆向方法。逆有限元模型将通过修正各向异性材料参数，最小化基于测量和预测位移的误差函数来求解。求解逆有限元模型是一项计算密集型任务，因为 (a) 有限元模型中存在大量自由度（通常 > 10,000）；(b) 逆问题中需要求解大量未知材料参数。逆解需要多处理器硬件和并行算法来提高逆问题的求解速度。我们建议通过基于子区域的域分解来实现逆算法的宏观并行化，以将每个处理器上一次处理的自由度数量减少到易于处理的水平。借助合适的多处理器硬件，所提出的算法可能会减少计算时间，而不会牺牲参数精度。

This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. The goal of the proposed project is to obtain anisotropic elastic properties of soft tissues from shear wave displacement measurements. This will be achieved by first simulating the forward problem with the finite element method i.e. shear wave propagation in an elastic material. Second, an inverse method will be setup based on measured displacements from elastography experiments and predicted displacements from the forward problem. The inverse finite element model will be solved by minimizing an error function based on measured and predicted displacements by correcting for the anisotropic material parameters. Solving the inverse finite element model is a computationally intensive task due to (a) large number of degrees of freedom (usually > 10,000) in the finite element model (b) the large number of unknown material parameters that need to be solved for in the inverse problem. The inverse solution will require multiprocessor hardware and parallel algorithms to enhance the speed of solution of the inverse problem. We propose to implement macro-level parallelization of the inverse algorithm through subzone based domain decomposition to reduce the number of degrees of freedom treated at one time on each processor to tractable levels. With suitable multiprocessor hardware the proposed algorithms potentially reduce the computational time without sacrificing parameter accuracy.