AV-Net: Deep learning for fully automated artery-vein classification in optical coherence tomography angiography

TL;DR

This paper introduces AV-Net, a deep learning model that automates artery-vein classification in OCTA images, combining multi-modal data and transfer learning to achieve high accuracy for clinical use.

Contribution

The study presents AV-Net, a novel fully convolutional network that integrates enface OCT and OCTA data with transfer learning for automated AV classification.

Findings

Achieved an average accuracy of 86.75% in AV classification.

Demonstrated the effectiveness of multi-modal data fusion in OCTA analysis.

Showcased the potential for fully automated AV analysis in clinical settings.

Abstract

This study is to demonstrate deep learning for automated artery-vein (AV) classification in optical coherence tomography angiography (OCTA). The AV-Net, a fully convolutional network (FCN) based on modified U-shaped CNN architecture, incorporates enface OCT and OCTA to differentiate arteries and veins. For the multi-modal training process, the enface OCT works as a near infrared fundus image to provide vessel intensity profiles, and the OCTA contains blood flow strength and vessel geometry features. A transfer learning process is also integrated to compensate for the limitation of available dataset size of OCTA, which is a relatively new imaging modality. By providing an average accuracy of 86.75%, the AV-Net promises a fully automated platform to foster clinical deployment of differential AV analysis in OCTA.