Martelli Fabrizio, Del Bianco Samuele, Spinelli Lorenzo, Cavalieri Stefano, Di Ninni Paola, Binzoni Tiziano, Jelzow Alexander, Macdonald Rainer, Wabnitz Heidrun
Università degli Studi di Firenze, Dipartimento di Fisica e Astronomia, Via G. Sansone 1, Sesto Fiorentino 50019, Firenze, Italy.
Istituto di Fisica Applicata Nello Carrara del Consiglio Nazionale delle Ricerche, Via Madonna del Piano 10, Sesto Fiorentino 50019, Italy.
J Biomed Opt. 2015 Nov;20(11):115001. doi: 10.1117/1.JBO.20.11.115001.
In this work, we have tested the optimal estimation (OE) algorithm for the reconstruction of the optical properties of a two-layered liquid tissue phantom from time-resolved single-distance measurements. The OE allows a priori information, in particular on the range of variation of fit parameters, to be included. The purpose of the present investigations was to compare the performance of OE with the Levenberg–Marquardt method for a geometry and real experimental conditions typically used to reconstruct the optical properties of biological tissues such as muscle and brain. The absorption coefficient of the layers was varied in a range of values typical for biological tissues. The reconstructions performed demonstrate the substantial improvements achievable with the OE provided a priori information is available. We note the extreme reliability, robustness, and accuracy of the retrieved absorption coefficient of the second layer obtained with the OE that was found for up to six fit parameters, with an error in the retrieved values of less than 10%. A priori information on fit parameters and fixed forward model parameters clearly improves robustness and accuracy of the inversion procedure.
在这项工作中,我们测试了最优估计(OE)算法,用于从时间分辨单距离测量中重建双层液体组织模型的光学特性。最优估计允许纳入先验信息,特别是关于拟合参数的变化范围。本研究的目的是在通常用于重建肌肉和大脑等生物组织光学特性的几何结构和实际实验条件下,比较最优估计与列文伯格 - 马夸特方法的性能。各层的吸收系数在生物组织典型的值范围内变化。所进行的重建表明,只要有先验信息,最优估计就能实现显著改进。我们注意到,对于多达六个拟合参数,用最优估计获得的第二层吸收系数的检索具有极高的可靠性、稳健性和准确性,检索值的误差小于10%。关于拟合参数和固定正向模型参数的先验信息明显提高了反演过程的稳健性和准确性。