Northwestern University (Bluhm Cardiovascular Institute)
nnU-Net-based segmentation network that detects and delineates stenotic lesions directly from X-ray coronary angiography frames, developed for the ARCADE (MICCAI 2023) stenosis-detection challenge. A companion model (YOLO-Angio, same team) handles vessel-tree segmentation; StenUNet focuses specifically on pixel-wise localization of stenotic regions. Placed 3rd overall among ARCADE challenge entrants with an F1 score of 0.5348 on the hold-out test set, within 0.0005 of the 2nd-place team.
Architecture
CNN (2D)
nnU-Net-based 2D U-Net segmentation network with custom preprocessing (multi-channel contrast enhancement) and postprocessing (small-segment removal) for pixel-wise stenosis detection
Framework
PyTorch
Added to catalog
2026-08-12
Apache 2.0
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.
Pixel-wise segmentation of stenotic lesions in X-ray coronary angiography frames (ARCADE 2023 challenge stenosis-detection task)