Retinal OCT Denoising with Pseudo-Multimodal Fusion Network

TL;DR

This paper introduces a learning-based fusion network that combines noisy OCT images with a pseudo-modality to effectively reduce speckle noise and enhance retinal layer contrast without requiring longer acquisition times.

Contribution

It proposes a novel pseudo-multimodal fusion approach that leverages self-fusion to improve OCT denoising while preserving fine structures and small vessels.

Findings

Significant noise suppression and contrast enhancement in retinal OCT images.

Improved structural similarity index from 0.559 to 0.576 compared to single modality.

Effective preservation of small blood vessels and tissue layers.

Abstract

Optical coherence tomography (OCT) is a prevalent imaging technique for retina. However, it is affected by multiplicative speckle noise that can degrade the visibility of essential anatomical structures, including blood vessels and tissue layers. Although averaging repeated B-scan frames can significantly improve the signal-to-noise-ratio (SNR), this requires longer acquisition time, which can introduce motion artifacts and cause discomfort to patients. In this study, we propose a learning-based method that exploits information from the single-frame noisy B-scan and a pseudo-modality that is created with the aid of the self-fusion method. The pseudo-modality provides good SNR for layers that are barely perceptible in the noisy B-scan but can over-smooth fine features such as small vessels. By using a fusion network, desired features from each modality can be combined, and the weight of their contribution is adjustable. Evaluated by intensity-based and structural metrics, the result shows that our method can effectively suppress the speckle noise and enhance the contrast between retina layers while the overall structure and small blood vessels are preserved. Compared to the single modality network, our method improves the structural similarity with low noise B-scan from 0.559 +\- 0.033 to 0.576 +\- 0.031.