MoE-based Feature Adapter for Prompt-free Binary Coronary Artery Segmentation in X-ray Angiography

Lin Xi; Yingliang Ma

cs.CVAug 25, 2026

MoE-based Feature Adapter for Prompt-free Binary Coronary Artery Segmentation in X-ray Angiography

Authors: Lin Xi, Yingliang Ma

Organizations: University College London, United Kingdom · University of East Anglia, United Kingdom · King’s College London, United Kingdom

Abstract

Accurate segmentation of coronary arteries in X-ray angiography videos is essential for quantitative coronary analysis and image-guided interventions. However, accurate segmentation remains challenging because coronary vessels are thin and exhibit low contrast, while the presence of catheters, guidewires, and complex anatomical background structures can further interfere with vessel delineation. Existing U-Net- and Transformer-based models provide strong baselines, but their shared feature-adaptation pathways may be insufficient for heterogeneous angiographic appearances. In this paper, we propose a prompt-free mixture-of-experts (MoE) feature adapter for binary coronary artery segmentation. Built upon parameter-efficient Vision Transformer adapters, the proposed method uses multiple lightweight experts with input-dependent top- $k$ routing to adaptively refine vessel-related features while limiting active computational cost. Experiments on MOSXAV and external evaluation on XACV show that the proposed method outperforms representative baselines and improves cross-dataset generalisation. These results suggest that MoE-based adapter learning is effective for robust coronary artery segmentation in X-ray angiography videos.

Figures & tables

Figure 1: Overall architecture of the proposed network.

Method	Validation Set				Test Set
Method	Dice $\uparrow$	IoU $\uparrow$	Precision $\uparrow$	Recall $\uparrow$	Dice $\uparrow$	IoU $\uparrow$	Precision $\uparrow$	Recall $\uparrow$
U-Net [ 4 ]	81.15	69.02	84.10	79.98	35.55	23.22	26.55	71.87
Attention U-Net [ 5 ]	81.76	69.86	86.50	79.11	38.68	25.85	29.43	70.27
nnU-Net [ 6 ]	83.14	72.03	92.01	73.58	45.76	28.59	40.33	73.04
Pseudo Diffusion [ 9 ]	80.82	66.89	82.36	72.63	34.78	22.97	24.26	68.11
AdapterSeg [ 17 ]	83.83	72.82	85.37	83.53	50.57	36.13	40.60	81.96

Table 1: Binary coronary artery segmentation results on the MOSXAV validation and test sets.

Method	Dice $\uparrow$	IoU $\uparrow$	Precision $\uparrow$	Recall $\uparrow$
U-Net [ 4 ]	65.58	49.62	69.32	62.99
Attention U-Net [ 5 ]	64.30	48.25	71.28	59.44
nnU-Net [ 6 ]	67.02	51.23	77.61	59.66
Pseudo Diffusion [ 9 ]	62.37	45.81	69.58	59.33
AdapterSeg [ 17 ]	68.41	52.76	76.20	62.95
MaskVSC [ 24 ]	60.56	44.37	69.20	54.65

Table 2: External binary vessel segmentation results on XACV.

Variant	Dice $\uparrow$	IoU $\uparrow$	Precision $\uparrow$	Recall $\uparrow$	Params
Baseline backbone	48.37	32.77	62.19	60.72	308M
+ Single adapter	83.83	72.82	85.37	83.53	325M
+ MoE adapter, all experts	84.25	73.91	85.08	84.88	356M
+ MoE adapter, top-1	83.76	73.19	85.13	83.47	356M
+ MoE adapter, top-k	84.75	74.05	85.22	85.97	356M

Table 3: Ablation study on MOSXAV.

Explore similar work

Multi-View Deformable Convolution Meets Visual Mamba for Coronary Artery Segmentation
Mar 23, 2026Xiaochan Yuan, Pai ZengCoronary Artery Segmentation 3D Medical Image Segmentation
Sichuan Agricultural University
Uncertainty-Aware Concept and Motion Segmentation for Semi-Supervised Angiography Videos
Mar 1, 2026Yu Luo, Guangyu Wei, Yangfan Li +2Semi-Supervised Medical Image Segmentation Coronary Artery Segmentation
School of Mathematical Sciences, Ocean University of China, China · School of Haide, Ocean University of China, China · School of Intelligent Science and Technology, Nanjing University, China +2
HTC-SGA Former: A Hybrid Transformer-CNN Network with Self-Guided Attention and a New Boundary-Weighted Adaptive Loss for Coronary DSA Vessel Segmentation
Jun 29, 2026Rayan Merghani Ahmed, Marwa Omer Mohammed Omer, Mohamed Elmanna +3Medical Imaging Coronary Angiography
Shenzhen Institutes of Advanced Technology (SIAT), Chinese Academy of Sciences (CAS), Shenzhen, 518055, China · University of Chinese Academy of Sciences (UCAS), Beijing, 101408, China · Department of Biomedical Engineering and Systems, Faculty of Engineering, Cairo University, Cairo, 12613, Egypt

Mar 23, 2026J/K move · Enter open · S save

Multi-View Deformable Convolution Meets Visual Mamba for Coronary Artery Segmentation

Xiaochan Yuan, Pai Zeng

Sichuan Agricultural University

Coronary Artery Segmentation 3D Medical Image Segmentation Deformable Convolution

Open paper AskPDF arXiv

Accurate segmentation of coronary arteries from computed tomography angiography (CTA) images is of paramount clinical importance for the diagnosis and treatment planning of cardiovascular diseases. However, coronary artery segmentation remains challenging due to the inherent multi-branching and slender tubular morphology of the vasculature, compounded by severe class imbalance between foreground vessels and background tissue. Conventional convolutional neural network (CNN)-based approaches struggle to capture long-range dependencies among spatially distant vascular structures, while Vision Transformer (ViT)-based methods incur prohibitive computational overhead that hinders deployment in resource-constrained clinical settings. Motivated by the recent success of state space models (SSMs) in efficiently modeling long-range sequential dependencies with linear complexity, we propose MDSVM-UNet, a novel two-stage coronary artery segmentation framework that synergistically integrates multidirectional snake convolution (MDSConv) with residual visual Mamba (RVM). In the encoding stage, we introduce MDSConv, a deformable convolution module that learns adaptive offsets along three orthogonal anatomical planes -- sagittal, coronal, and axial -- thereby enabling comprehensive multi-view feature fusion that faithfully captures the elongated and tortuous geometry of coronary vessels. In the decoding stage, we design an RVM-based upsampling decoder block that leverages selective state space mechanisms to model inter-slice long-range dependencies while preserving linear computational complexity. Furthermore, we propose a progressive two-stage segmentation strategy: the first stage performs coarse whole-image segmentation to guide intelligent block extraction, while the second stage conducts fine-grained block-level segmentation to recover vascular details and suppress false positives..

Abstract

Explore similar work

Multi-View Deformable Convolution Meets Visual Mamba for Coronary Artery Segmentation

Uncertainty-Aware Concept and Motion Segmentation for Semi-Supervised Angiography Videos

HTC-SGA Former: A Hybrid Transformer-CNN Network with Self-Guided Attention and a New Boundary-Weighted Adaptive Loss for Coronary DSA Vessel Segmentation