]> genBornMat - compute forward problem

Compute Forward Problem

genBornMat() is a general routine for solving the forward problem. Based on the Method argument, the geometry in the Medium structure, and the contents of the measurement list, genBornMat() figures out what the appropriate low-level routine to call is and makes the appropriate call for the user.

Function Summary

Syntax: [Phi0, A] = genBornMat(SD, Medium, MeasList, Method, muVec);
Inputs: SD SD structure
Medium Medium structure
MeasList Measurement List that corresponds to experimental data
Method Which method to use (Born, Rytov, etc.)
muVec Flags to indicate optical perturbations (see below)
Outputs: Phi0 The fluence in a homogeneous medium
A The forward matrix

Detailed Descriptions

The muVec argument is a two-parameter vector. The first element (muVec(1)) indicates whether A should include terms for calculating absorption perturbations, the second term muVec(2) is for scattering perturbations. If the flag is non-zero, the forward matrix includes those elements, otherwise they are not calculated. Either muVec(1), muVec(2) (or both) should be non-zero.

The forward matrix has a block form. To save space, a sparse matrix representation is used for all but the simplest instances of the forward matrix A. The outermost block is the type of perturbation (if both are used); absorption first, then scattering. Within that block, the next set of blocks are by wavelength (so that the matrix will work correctly with perturbations at multiple wavelengths). Space is left for all the wavelengths in SD, regardless of whether those wavelengths appear in the measurement list. Finally, within each wavelength is a dense matrix mapping perturbations at that wavelength into the "scattered" field.

genBornMat() is mostly just a wrapper function that calls the appropriate low-level routines based on the contents of the measurement list and the Method argument. For frequency-domain imagers, the conversion from the "Born" to "Rytov" formulation is also handled by genBornMat().