Time Resolved Diffuse Optical Tomography model

Abstract

Time Resolved Diffuse Optical Tomography (TRDOT) modality is presented. Images were reconstructed for four different scenarios. 64-source and 64-detector bifurcated positions were used for simulation model. Inclusion was buried in different coordinate locations. Time-dependent diffusion equation for photon-tissue interactions were used to create the forward model. Mostly used TRDOT devices have pulsed-laser source and photo-multiplier tube (PMT) photodetectors. In addition to pulsed laser source, continuous wave (CW) sources would be used for TRDOT imaging. In general usage, solid-state diode pumped pulsed-lasers and driving units constitute source module. Solid-state lasers are made of Titanium Sapphire (TiSa). These laser sources are difficult to use, modify, maintain, and they are also expensive. Instead of using expensive pulsed laser sources, cheap electronic-based pulsed-laser driving circuit will be implemented. Pseudomorphic high electron mobility transistors (pHEMTs) as switching elements will be used at both pulsed-laser and photodiode current readout sides. Hence, this work is taking its motivation from the next-generation electronic-based device instrument, creating the time dependent forward model as its focus. In this work, it was seen that selection of time intervals are important parameters to reconstruct the hidden images inside the homogeneous tissue. Inclusion was buried inside the homogeneous tissue model, and corresponding time modes have been selected to extract the hidden inclusion. In this work, time modes were analyzed for the TRDOT device. It has been realized that different time mode forward model weight functions should be used for different depth layers to reconstruct the hidden inclusion correctly. Hence, different time mode forward model weight matrix functions were generated from time 20 ps to 400 ps by 20 ps steps, and from time 2 ps to 40 ps by 2 ps steps. In the mathematical inverse problem solution algorithm, these time clusters were used to create two different weight matrixes.