DeNVeR: Deformable Neural Vessel Representations for Unsupervised Video Vessel Segmentation

2024-06-03

arXiv.org

AI Summary

This paper addresses the unsupervised vascular segmentation challenge in X-ray angiographic videos, where pixel-level annotations are unavailable. We propose a novel optical-flow-guided, layer-separation-driven test-time adaptation method. Our key contributions are: (1) a layer-separation bootstrapping strategy coupled with parallel vascular motion loss to decouple static vascular anatomy from dynamic motion components; (2) Eulerian motion field modeling of vascular deformation, integrated with neural radiance field-based deformation representation to enhance temporal consistency; and (3) the first publicly available coronary angiography video dataset—XACV—with fine-grained manual annotations. Evaluated on XACV and CADICA, our method achieves state-of-the-art performance in segmentation accuracy, inter-frame consistency, and cross-domain generalization, significantly outperforming existing unsupervised approaches.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Calibration & Uncertainty Quantification
• Search and Optimization: Non-convex Optimization

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

This paper presents Deformable Neural Vessel Representations (DeNVeR), an unsupervised approach for vessel segmentation in X-ray angiography videos without annotated ground truth. DeNVeR utilizes optical flow and layer separation techniques, enhancing segmentation accuracy and adaptability through test-time training. Key contributions include a novel layer separation bootstrapping technique, a parallel vessel motion loss, and the integration of Eulerian motion fields for modeling complex vessel dynamics. A significant component of this research is the introduction of the XACV dataset, the first X-ray angiography coronary video dataset with high-quality, manually labeled segmentation ground truth. Extensive evaluations on both XACV and CADICA datasets demonstrate that DeNVeR outperforms current state-of-the-art methods in vessel segmentation accuracy and generalization capability while maintaining temporal coherency. See our project page for video results at https://kirito878.github.io/DeNVeR/.

Problem

• Unsupervised vessel segmentation in X-ray videos
• Modeling complex vessel dynamics without annotations
• Improving accuracy and generalization in segmentation

Innovation

• Unsupervised vessel segmentation using optical flow
• Layer separation bootstrapping for enhanced accuracy
• Eulerian motion fields for complex vessel dynamics