Ocean University of China / Shandong University
Segments coronary vessels from invasive X-ray angiography images and automatically quantifies the degree of stenosis along the extracted centerlines. Combines MedSAM, a Segment-Anything-style vision model, with a Mamba-based VM-UNet segmentation branch for efficient long-range feature modeling. Trained and evaluated on the ARCADE, DCA1, and GH angiography datasets by researchers at Ocean University of China and Shandong University.
Architecture
Hybrid
Two-branch: MedSAM (ViT-based Segment Anything Model) feature extractor feeding a VM-UNet (Vision-Mamba state-space UNet) segmentation branch
Framework
PyTorch
Added to catalog
2026-07-10
MIT
License for model weights only. Associated code may be licensed seperately, check code source for specific terms.
Automatic Region-based Coronary Artery Disease diagnostics dataset of invasive X-ray coronary angiography images with vessel segmentation annotations.
Coronary vessel segmentation (mixed dataset)
Coronary vessel segmentation (ARCADE only)
Stenosis detection