Song Xiaolei, Yi Ji, Bai Jing
Department of Biomedical Engineering, Tsinghua University, Beijing 100084, China.
Int J Biomed Imaging. 2006;2006:70839. doi: 10.1155/IJBI/2006/70839. Epub 2006 Dec 14.
Based on an independent forward model in fluorescent tomography, a parallel reconstructed scheme for inhomogeneous mediums with unknown absorption property is proposed in this paper. The method considers the two diffusion equations as separately describing the propagation of excited light in tissues with and without fluorescent probes inside. Then the concentration of fluorophores is obtained directly through the difference between two estimations of absorption coefficient which can be parallel inversed. In this way, the multiparameter estimation problem in fluorescent tomography is transformed into two independent single-coefficient determined schemes of diffusion optical tomography (DOT). Any algorithms proved to be efficient and effective in DOT can be directly applied here. In this study the absorption property is estimated from the independent diffusion equations by a gradient-based optimization method with finite element method (FEM) solving the forward model. Simulation results of three representative occasions show that the reconstructed method can well estimate fluorescent property and tissue absorption distribution.
基于荧光层析成像中的独立正向模型，本文提出了一种针对吸收特性未知的非均匀介质的并行重建方案。该方法将两个扩散方程分别视为描述激发光在有荧光探针和无荧光探针的组织中传播的方程。然后，通过两个可并行反演的吸收系数估计值之间的差异直接获得荧光团的浓度。通过这种方式，荧光层析成像中的多参数估计问题被转化为扩散光学层析成像（DOT）的两个独立的单系数确定方案。任何在DOT中被证明是高效有效的算法都可以直接应用于此。在本研究中，通过基于梯度的优化方法从独立的扩散方程估计吸收特性，并使用有限元方法（FEM）求解正向模型。三个代表性实例的模拟结果表明，该重建方法能够很好地估计荧光特性和组织吸收分布。
