Abstract
Real-time X-ray fluoroscopy is the primary guidance modality in interventional cardiovascular procedures. While continuous lowdose fluoroscopy is safe, it suffers from serious problems such as noise, blurring, and low contrast. Therefore, clinicians intermittently acquire high-dose images, but these high-dose images, although high-resolution, are out of sync in time. In this study, we propose a physics-consistent and region-aware temporal enhancement framework tailored for interventional fluoroscopy. To address degradation and supervision mismatch in clinical workflows of fluoroscopy acquisition, we introduce a dual-path degradation modeling strategy that combines analytic simulation based on the system Point Spread Function (PSF) to model physics-consistent blur, together with diffusion-driven hierarchical degradation synthesis, enabling robust pseudo-paired training under realistic imaging conditions. To improve temporal stability in low-contrast and motion-unstable sequences, we integrate frequency-aware operator modeling with efficient temporal propagation. A region-aware optimization scheme further prioritizes anatomically critical structures, including cardiac boundaries, spine contours, and catheter-related regions. Experiments under both synthetic paired and real-world unpaired settings demonstrate consistent improvements in reconstruction accuracy and temporal stability while maintaining real-time feasibility, suggesting the potential of the proposed framework for improving interventional procedures.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/DT4H_031.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=lvsiVWDX9d
BibTex
@InProceedings{ZhaYus_RegionAware_MICCAISAT2026,
author = { Zhang, Yusongjia AND Wang, Xingyao AND Qi, Xiuyu AND Cui, Chang AND Ju, Weizhu AND Li, Lei},
title = { { Region-Aware Temporal Super-Resolution for Interventional X-ray Imaging } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17275},
month = {pending},
page = {pending}
}