Time-domain diffuse optical tomography using recursive direct method of calculating Jacobian at selected temporal points

Mohamed A. Naser, M. J. Deen

Research output: Contribution to journalArticlepeer-review

8 Scopus citations

Abstract

An algorithm for time-domain diffuse optical tomography basedon the resolution of the timedomain diffusion equation using the finite element method has been developed. An efficient direct method including are cursive approach has been used toobtain the lightfluence derivatives with respect to tissue optical properties at precise selected points on the temporal profile resulting in a considerable savings in computation time and memory. The algorithm reconstructs the tissue optical properties in a permissible region or a region-of-interest and the input data for reconstruction comprises selections of points on the temporal curve of the measured pulse. The optical properties have been reconstructed by solvingan iterative normalized minimization problem. The algorithm has been appliedto a three-dimensional simplified model of a new born baby head and to a three-dimensional model of the mouse (MOBY) for a small animal model. The computation speed and memory usage of the algorithm have been compared with that of other techniques based on continuous wave and frequency domain representations. The effects of using different sizes of time steps and number of time steps onthe reconstruction accuracy and the computation time have been reported.

Original languageEnglish (US)
Article number045207
JournalBiomedical Physics and Engineering Express
Volume1
Issue number4
DOIs
StatePublished - Nov 18 2015
Externally publishedYes

Keywords

  • Inverse problems
  • Light propagation intissues
  • Medical and biological imaging
  • Time resolved imaging
  • Tomography

ASJC Scopus subject areas

  • General Nursing

Fingerprint

Dive into the research topics of 'Time-domain diffuse optical tomography using recursive direct method of calculating Jacobian at selected temporal points'. Together they form a unique fingerprint.

Cite this