Boundary-Aware Vision Transformer for Angiography Vascular Network Segmentation

2025-06-15

AI Summary

Segmenting slender, low-contrast coronary arteries in angiography is challenging due to ambiguous boundaries and topological discontinuities. This paper proposes a boundary-aware segmentation model based solely on a Vision Transformer (ViT) architecture—eliminating all CNN components to preserve a lightweight, fully transformer-based design that supports scalable pretraining and transfer learning. The key innovation is an edge-aware loss function that explicitly guides the ViT to learn precise vascular boundaries. Evaluated on the DCA-1 dataset, the model achieves state-of-the-art performance across critical metrics—including Dice score, 95th-percentile Hausdorff distance (HD95), and average surface distance (ASD)—significantly outperforming both conventional CNNs and hybrid CNN-ViT approaches. It delivers clinically viable segmentation with superior accuracy and topological continuity.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Learning on the Edge & Model Compression
• Constraint Satisfaction and Optimization: Satisfiability Modulo Theories

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Accurate segmentation of vascular structures in coronary angiography remains a core challenge in medical image analysis due to the complexity of elongated, thin, and low-contrast vessels. Classical convolutional neural networks (CNNs) often fail to preserve topological continuity, while recent Vision Transformer (ViT)-based models, although strong in global context modeling, lack precise boundary awareness. In this work, we introduce BAVT, a Boundary-Aware Vision Transformer, a ViT-based architecture enhanced with an edge-aware loss that explicitly guides the segmentation toward fine-grained vascular boundaries. Unlike hybrid transformer-CNN models, BAVT retains a minimal, scalable structure that is fully compatible with large-scale vision foundation model (VFM) pretraining. We validate our approach on the DCA-1 coronary angiography dataset, where BAVT achieves superior performance across medical image segmentation metrics outperforming both CNN and hybrid baselines. These results demonstrate the effectiveness of combining plain ViT encoders with boundary-aware supervision for clinical-grade vascular segmentation.

Problem

• Accurate segmentation of complex coronary angiography vessels
• Preserving topological continuity in vascular network segmentation
• Enhancing boundary awareness in Vision Transformer models

Innovation

• Boundary-aware ViT for vascular segmentation
• Edge-aware loss enhances boundary precision
• Pure ViT structure compatible with VFM