CVAI Catalog

·

View Catalog

CM-UNet (Contrastive Masked UNet)

EPFL / Lausanne University Hospital (CHUV)

Coronary angiography

Filter catalog by Modality:
Invasive Coronary & Intracoronary Imaging

Coronary artery segmentation / anatomy

Filter catalog by Disease / Trait:
Coronary & Ischemic Disease

Segmentation

Filter catalog by Task Type:
Segmentation & Detection

Hybrid

Filter catalog by Architecture:
Hybrid / Multi-branch

PyTorch

Filter catalog by Framework:
PyTorch

Self-supervised deep learning model for coronary artery segmentation from invasive X-ray coronary angiography (ICA), designed to reduce reliance on large annotated datasets. CM-UNet combines a Contrastive Masked Autoencoder (CMAE) with a UNet backbone: an online encoder-decoder branch reconstructs masked image patches while a momentum branch produces contrastive embeddings, jointly pretraining the network on unannotated angiography images before fine-tuning on a small labeled set. Fine-tuning with only 18 annotated images (instead of 500) led to just a 15.2% drop in Dice score, versus a 46.5% drop for baseline models trained without this self-supervised pretraining -- demonstrating strong label efficiency for coronary segmentation.

memory Specifications

category

Architecture

Hybrid

Contrastive Masked Autoencoder (CMAE) combined with a UNet backbone: an online encoder-decoder reconstructs masked image patches while a momentum-updated branch generates contrastive embeddings for self-supervised pretraining, followed by supervised fine-tuning on limited annotated coronary angiography data

code

Framework

PyTorch

calendar_month

Added to catalog

2026-08-14

description Publication

CM-UNet: A Self-Supervised Learning-Based Model for Coronary Artery Segmentation in X-Ray Angiography open_in_new

Challier C, Sun X, Mahendiran T, Senouf O, De Bruyne B, Auberson D, Muller O, Fournier S, Frossard P, Abbe E, Thanou D

IEEE EMBC 2025 · 2025 · original paper

DOI: 10.1109/EMBC58623.2025.11253755

database Training & evaluation data

Unannotated X-ray coronary angiography corpus (CM-UNet)

train

Unannotated invasive X-ray coronary angiography images used for self-supervised contrastive-masked pretraining, plus a limited annotated set (up to 500 images, as few as 18 used in label-efficiency experiments) for fine-tuning.

science Capabilities & performance

Segmentation of coronary arteries from X-ray angiography images, using self-supervised pretraining plus limited-label fine-tuning

Segmentation Coronary artery segmentation / anatomy