VAFO-Loss: VAscular Feature Optimised Loss Function for Retinal Artery/Vein Segmentation
This work addresses a domain-specific issue in medical imaging by improving segmentation for clinical applications, though it is incremental as it builds on existing segmentation networks.
The paper tackles the problem of inaccurate vascular feature estimation in retinal artery/vein segmentation by introducing VAFO-Loss, a novel loss function that integrates clinical features like vessel density and fractal dimension, resulting in improved segmentation metrics and quantitative enhancement in stroke incidence prediction.
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.