VAFO-Loss: VAscular Feature Optimised Loss Function for Retinal Artery/Vein Segmentation

TL;DR

This paper introduces VAFO-Loss, a novel loss function that incorporates vascular features into retinal vessel segmentation networks, improving both feature estimation and downstream clinical prediction accuracy.

Contribution

The paper proposes VAFO-Loss, a new vascular feature optimized loss function that enhances retinal vessel segmentation and vascular feature extraction.

Findings

Improved vascular feature estimation with VAFO-Loss.

Enhanced downstream stroke prediction accuracy.

Statistically significant segmentation metric improvements.

Abstract

Estimating clinically-relevant vascular features following vessel segmentation is a standard pipeline for retinal vessel analysis, which provides potential ocular biomarkers for both ophthalmic disease and systemic disease. In this work, we integrate these clinical features into a novel vascular feature optimised loss function (VAFO-Loss), in order to regularise networks to produce segmentation maps, with which more accurate vascular features can be derived. Two common vascular features, vessel density and fractal dimension, are identified to be sensitive to intra-segment misclassification, which is a well-recognised problem in multi-class artery/vein segmentation particularly hindering the estimation of these vascular features. Thus we encode these two features into VAFO-Loss. We first show that incorporating our end-to-end VAFO-Loss in standard segmentation networks indeed improves vascular feature estimation, yielding quantitative improvement in stroke incidence prediction, a clinical downstream task. We also report a technically interesting finding that the trained segmentation network, albeit biased by the feature optimised loss VAFO-Loss, shows statistically significant improvement in segmentation metrics, compared to those trained with other state-of-the-art segmentation losses.