Imaging-formulation-based numerical speckle reduction for optical coherence tomography

TL;DR

This paper introduces a numerical speckle reduction method for OCT that effectively suppresses speckle noise using a single volumetric acquisition, preserving resolution and revealing microstructures.

Contribution

The study presents a novel speckle reduction technique based on the dispersed scatterer model and imaging formulation, outperforming conventional methods without hardware changes.

Findings

Significant speckle suppression demonstrated in phantom and biological samples.

Preservation of lateral resolution confirmed through phantom measurements.

Enhanced visualization of microstructures like necrotic regions in spheroids.

Abstract

Speckle is an intrinsic pattern in optical coherence tomography (OCT) that obscures fine image features and degrades effective resolution. In this study, we propose a numerical speckle reduction method based on the dispersed scatterer model and the imaging formulation of OCT. Utilizing the shifted-complex-conjugate-product, the proposed method digitally modulates speckle patterns by shifting the complex en face OCT signal and averaging the resulting real-part images. This approach allows for effective speckle suppression using a single volumetric acquisition without additional hardware modifications. OCT point spread function phantom measurement demonstrated lateral resolution preservation of the proposed method. We validated the method using a custom-built full-field swept-source OCT system on human breast adenocarcinoma spheroids and a zebrafish eye. Quantitative evaluations using the contrast-to-noise ratio and equivalent number of looks demonstrated that the proposed method significantly outperforms conventional frame-averaging techniques. The speckle-reduced images revealed microstructures previously obscured by speckle, such as necrotic regions in spheroids, while preserving the original image sharpness and resolution.