Computational Methods for Spectroscopic Optical Coherence Tomography

This thesis develops computational methods for extracting quantitative spectroscopic information from optical coherence tomography data. Three primary contributions are presented: (1) a physics-constrained neural network for depth-resolved spectral estimation, (2) a Jones matrix simulation framework for polarization-sensitive OCT system design, and (3) a spectroscopic intravascular OCT system for coronary artery tissue characterization. Each contribution is validated against established methods and experimental data.