Optical Tomography With a Least Square Finite Element Formulation of the Collimated Irradiation in Frequency Domain

Author(s):  
Olivier Balima ◽  
Joan Boulanger ◽  
Andre´ Charette ◽  
Daniel Marceau

This paper presents a numerical study of optical tomography in frequency domain for the reconstruction of optical properties of scattering and absorbing media with collimated irradiation light sources. The forward model is a least square finite element formulation of the collimated irradiation problem where the intensity is separated into its collimated and scattered parts. This model does not use any empirical stabilization and moreover the collimated source direction is taken into account. The inversion uses a gradient type minimization method where the gradient is computed through an adjoint formulation. Scaling is used to avoid numerical round errors, as the output readings at detectors are very low. Numerical reconstructions of optical properties of absorbing and scattering media with simulated data (noised and noise-free) are achieved in a complex geometry with satisfactory results. The results show that complex geometries are well handled with the proposed method.

2011 ◽  
Vol 2011 ◽  
pp. 1-11 ◽  
Author(s):  
Martin Schweiger

We introduce a GPU-accelerated finite element forward solver for the computation of light transport in scattering media. The forward model is the computationally most expensive component of iterative methods for image reconstruction in diffuse optical tomography, and performance optimisation of the forward solver is therefore crucial for improving the efficiency of the solution of the inverse problem. The GPU forward solver uses a CUDA implementation that evaluates on the graphics hardware the sparse linear system arising in the finite element formulation of the diffusion equation. We present solutions for both time-domain and frequency-domain problems. A comparison with a CPU-based implementation shows significant performance gains of the graphics accelerated solution, with improvements of approximately a factor of 10 for double-precision computations, and factors beyond 20 for single-precision computations. The gains are also shown to be dependent on the mesh complexity, where the largest gains are achieved for high mesh resolutions.


Sign in / Sign up

Export Citation Format

Share Document