LUNet: Deep Learning for the Segmentation of Arterioles and Venules in High Resolution Fundus Images

TL;DR

LUNet is a novel deep learning model designed for precise segmentation of retinal arterioles and venules in high-resolution fundus images, aiding cardiovascular disease diagnosis through improved microvasculature analysis.

Contribution

The paper introduces LUNet, a new deep learning architecture with a unique dilated convolutional block and high-resolution refinement, along with a new annotated dataset for retinal vessel segmentation.

Findings

LUNet outperforms existing segmentation algorithms on multiple datasets.

The model maintains high accuracy across diverse ethnicities and health conditions.

The created dataset is publicly available for further research.

Abstract

The retina is the only part of the human body in which blood vessels can be accessed non-invasively using imaging techniques such as digital fundus images (DFI). The spatial distribution of the retinal microvasculature may change with cardiovascular diseases and thus the eyes may be regarded as a window to our hearts. Computerized segmentation of the retinal arterioles and venules (A/V) is essential for automated microvasculature analysis. Using active learning, we created a new DFI dataset containing 240 crowd-sourced manual A/V segmentations performed by fifteen medical students and reviewed by an ophthalmologist, and developed LUNet, a novel deep learning architecture for high resolution A/V segmentation. LUNet architecture includes a double dilated convolutional block that aims to enhance the receptive field of the model and reduce its parameter count. Furthermore, LUNet has a long tail that operates at high resolution to refine the segmentation. The custom loss function emphasizes the continuity of the blood vessels. LUNet is shown to significantly outperform two state-of-the-art segmentation algorithms on the local test set as well as on four external test sets simulating distribution shifts across ethnicity, comorbidities, and annotators. We make the newly created dataset open access (upon publication).