Real-time diffuse optical tomography using reduced-order light propagation models based on a priori anatomical and functional information

This paper proposes a new fast 3D image reconstruction algorithm for Diffuse Optical Tomography using reduced order polynomial mappings from the space of optical tissue parameters into the space of flux measurements at the detector locations. The polynomial mappings are constructed through an iterative estimation process involving structure detection, parameter estimation and cross-validation using data generated by simulating a diffusion approximation of the radiative transfer equation incorporating a priori anatomical and functional information provided by MR scans and prior psychological evidence. Numerical simulation studies demonstrate that reconstructed images are remarkably similar in quality as those obtained using the standard approach, but obtained at a fraction of the time.