Classification of Neural Tissue Types for an Endoscopic Full-field Optical Coherence Tomography Imaging System

Optical coherence tomography is a volumetric imaging technique that can differentiate brain regions with micron-scale resolution. Our group has previously developed an endoscopic full-field swept-source OCT system with the aim to provide real-time surgical guidance for the implantation of electrodes for deep brain stimulation. This thesis builds on our prior work. Due to the limitations in the resolution and contrast due to scattering from the tissue, to identify different types of neural tissues, we developed an image classification algorithm that uses texture features and optical attenuation. The algorithm could classify three different brain regions in the ex vivo data set with a test accuracy of 70%. Also, we performed a comparative analysis on the optical properties of the ex vivo and in vivo data sets to provide directions on how to achieve high accuracy for in vivo data.M.A.S