Arbitrary-Scale Super-Resolution for Non-Contrast Coronary MRA via Wavelet-Laplacian Enhanced Implicit Network

Non-contrast coronary magnetic resonance angiography (NC-CMRA) provides a radiation-free and contrast-free approach for coronary assessment, but its diagnostic value is limited by insufficient spatial resolution. We propose an arbitrary-scale super-resolution framework that integrates a physiology-aware degradation model with a Wavelet-Laplacian Enhanced Implicit Network (WLEI-Net). The degradation model sequentially represents heart-rate-variability-driven asymmetric cardiac motion, Ornstein-Uhlenbeck respiratory translation, cardiopulmonary coupling, band-limited k-space noise, and Rician magnitude formation, and is designed to better represent the coupled physiological and k-space effects of clinical acquisitions. WLEI-Net combines SwinIR feature extraction, directional wavelet subband enhancement with channel attention, noise-adaptive Laplacian edge compensation, and a hybrid Charbonnier-wavelet-Sobel loss. Using 380 NC-CMRA images from 12 patients with a fixed patient-level split, the arbitrary-scale models were trained at 4.0× and evaluated at 2.0×, 2.5×, 3.0×, 3.5×, and 4.0×, with the four other evaluated scales unseen during training. WLEI-Net achieved the best mean performance for all six metrics at all five scales, and all comparisons against the four baselines remained significant after Benjamini-Hochberg false-discovery-rate correction (BH-adjusted p ≤ 0.012). In a blinded two-reader evaluation of 57 test images at 2.5× and 4.0×, WLEI-Net achieved the highest mean score for artifact/noise suppression, vascular edge integrity, and small-branch visibility, with ICC(2,2)=0.817-0.901. These results support internal reconstruction and reader-perceived image-quality benefits.

Authors

Institutions

Publication Details

Journal
Biomedical Physics & Engineering Express
Published
2026-11-17
DOI
https://doi.org/10.1088/2057-1976/aec2f3
Primary Topic
Advanced Image Processing Techniques
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Arbitrary-Scale Super-Resolution for Non-Contrast Coronary MRA via Wavelet-Laplacian Enhanced Implicit Network

Funan Xiao, chuanbing wang, 张闰, Xiuhan Li et al.
2 links
Biomedical Physics & Engineering Express
Advanced Image Processing Techniques
article

Arbitrary-Scale Super-Resolution for Non-Contrast Coronary MRA via Wavelet-Laplacian Enhanced Implicit Network

Funan Xiao, chuanbing wang, 张闰, Xiuhan Li, Yue Jiang, Da Cao, Yue Yin
article en

Abstract

Non-contrast coronary magnetic resonance angiography (NC-CMRA) provides a radiation-free and contrast-free approach for coronary assessment, but its diagnostic value is limited by insufficient spatial resolution. We propose an arbitrary-scale super-resolution framework that integrates a physiology-aware degradation model with a Wavelet-Laplacian Enhanced Implicit Network (WLEI-Net). The degradation model sequentially represents heart-rate-variability-driven asymmetric cardiac motion, Ornstein-Uhlenbeck respiratory translation, cardiopulmonary coupling, band-limited k-space noise, and Rician magnitude formation, and is designed to better represent the coupled physiological and k-space effects of clinical acquisitions. WLEI-Net combines SwinIR feature extraction, directional wavelet subband enhancement with channel attention, noise-adaptive Laplacian edge compensation, and a hybrid Charbonnier-wavelet-Sobel loss. Using 380 NC-CMRA images from 12 patients with a fixed patient-level split, the arbitrary-scale models were trained at 4.0× and evaluated at 2.0×, 2.5×, 3.0×, 3.5×, and 4.0×, with the four other evaluated scales unseen during training. WLEI-Net achieved the best mean performance for all six metrics at all five scales, and all comparisons against the four baselines remained significant after Benjamini-Hochberg false-discovery-rate correction (BH-adjusted p ≤ 0.012). In a blinded two-reader evaluation of 57 test images at 2.5× and 4.0×, WLEI-Net achieved the highest mean score for artifact/noise suppression, vascular edge integrity, and small-branch visibility, with ICC(2,2)=0.817-0.901. These results support internal reconstruction and reader-perceived image-quality benefits.

Biomedical Physics & Engineering Express
Nanjing University of Chinese Medicine (CN), Zhongda Hospital Southeast University (CN), Jiangsu Province Hospital (CN), Nanjing Medical University (CN)
Good health and well-being
Advanced Image Processing Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon