Statistical Estimation of Ballistic Signal in Visible Light OCT Based on Random Matrix Description

TL;DR

This paper introduces a random matrix model for vis-OCT signals that separates single and multiple scattering components, enhancing image quality and SNR in human fundus imaging.

Contribution

The study develops a Wishart random matrix model to distinguish scattering components in vis-OCT, improving imaging clarity and noise suppression.

Findings

Significantly improved imaging details of vascular structures.

Increased SNR by over 10dB in human fundus images.

Enhanced layer structure visibility with blood scattering suppression.

Abstract

Visible light optical coherence tomography (vis-OCT) provides a unique tool for imaging the structure and oxygen metabolism in tissues. However, since it works in the spectral domain, vis-OCT still suffers from noises due to the multiple scatterings, e.g. for imaging the human fundus. In this study, we modeled the OCT signals as a hybrid of single and multiple scattering components using Wishart random matrix description, with which the single scattering component thus can be separated out using the low-rank characteristics of the matrix. The model was validated using Monte Carlo simulation. We further demonstrated that this model could significantly improve the imaging performances in human fundus, showing more details of the vascular structure than the current vis-OCT and an increase of signal-to-noise ratio (SNR) up to more than 10dB. The layer structure of the retina can be better revealed with more than 3dB suppression of the blood scattering in OCT signals.