DynaFreq-SR: A Dynamic Spatially-Adaptive Network with Frequency-Domain Reconstruction for Chest X-ray Super-Resolution

Yuanxin Su; Thaiyal Bagam; Hailing Li; Zhijie Xu; Yangbin Chen; Haiyang Zhang; Yihong Wang

doi:10.1016/j.ifacol.2025.12.552

DynaFreq-SR: A Dynamic Spatially-Adaptive Network with Frequency-Domain Reconstruction for Chest X-ray Super-Resolution

Yuanxin Su^*
, Thaiyal Bagam^*
, Hailing Li
, Zhijie Xu
, Yangbin Chen
, Haiyang Zhang
, Yihong Wang

^*Corresponding author for this work

Xi'an Jiaotong-Liverpool University

Xi'an Jiaotong-Liverpool University
Communication University of China
Department of Computer Science and Software Engineering

Research output: Contribution to journal › Conference article › peer-review

Abstract

This paper proposes a deep CNN for single-image super-resolution of chest X-rays, integrating multi-scale context modules, dynamic spatially adaptive convolutions, and Fourier-based upsampling to recover fine diagnostic details from low-resolution inputs. Experiments on benchmark datasets show superior restoration quality and efficiency, demonstrating its promise as a clinical cognitive support tool to enhance radiologists' visual perception and diagnostic accuracy.

Original language	English
Pages (from-to)	649-654
Number of pages	6
Journal	IFAC-PapersOnLine
Volume	59
Issue number	35
DOIs	https://doi.org/10.1016/j.ifacol.2025.12.552
Publication status	Published - 2025
Event	16th IFAC Symposium on Analysis, Design and Evaluation of Human-Machine Systems, HMS 2025 - Beijing, China Duration: 18 Nov 2025 → 21 Nov 2025

Keywords

Chest X-ray
Deep Learning
Human-Machine System
Medical Imaging
Super Resolution

Access to Document

10.1016/j.ifacol.2025.12.552

Cite this

@article{d0caed51e2b44b8f8fcde656fbab518f,

title = "DynaFreq-SR: A Dynamic Spatially-Adaptive Network with Frequency-Domain Reconstruction for Chest X-ray Super-Resolution",

abstract = "This paper proposes a deep CNN for single-image super-resolution of chest X-rays, integrating multi-scale context modules, dynamic spatially adaptive convolutions, and Fourier-based upsampling to recover fine diagnostic details from low-resolution inputs. Experiments on benchmark datasets show superior restoration quality and efficiency, demonstrating its promise as a clinical cognitive support tool to enhance radiologists' visual perception and diagnostic accuracy.",

keywords = "Chest X-ray, Deep Learning, Human-Machine System, Medical Imaging, Super Resolution",

author = "Yuanxin Su and Thaiyal Bagam and Hailing Li and Zhijie Xu and Yangbin Chen and Haiyang Zhang and Yihong Wang",

note = "Publisher Copyright: {\textcopyright} 2025 Elsevier B.V.. All rights reserved.; 16th IFAC Symposium on Analysis, Design and Evaluation of Human-Machine Systems, HMS 2025 ; Conference date: 18-11-2025 Through 21-11-2025",

year = "2025",

doi = "10.1016/j.ifacol.2025.12.552",

language = "English",

volume = "59",

pages = "649--654",

journal = "IFAC-PapersOnLine",

issn = "2405-8971",

number = "35",

}

TY - JOUR

T1 - DynaFreq-SR

T2 - 16th IFAC Symposium on Analysis, Design and Evaluation of Human-Machine Systems, HMS 2025

AU - Su, Yuanxin

AU - Bagam, Thaiyal

AU - Li, Hailing

AU - Xu, Zhijie

AU - Chen, Yangbin

AU - Zhang, Haiyang

AU - Wang, Yihong

PY - 2025

Y1 - 2025

N2 - This paper proposes a deep CNN for single-image super-resolution of chest X-rays, integrating multi-scale context modules, dynamic spatially adaptive convolutions, and Fourier-based upsampling to recover fine diagnostic details from low-resolution inputs. Experiments on benchmark datasets show superior restoration quality and efficiency, demonstrating its promise as a clinical cognitive support tool to enhance radiologists' visual perception and diagnostic accuracy.

AB - This paper proposes a deep CNN for single-image super-resolution of chest X-rays, integrating multi-scale context modules, dynamic spatially adaptive convolutions, and Fourier-based upsampling to recover fine diagnostic details from low-resolution inputs. Experiments on benchmark datasets show superior restoration quality and efficiency, demonstrating its promise as a clinical cognitive support tool to enhance radiologists' visual perception and diagnostic accuracy.

KW - Chest X-ray

KW - Deep Learning

KW - Human-Machine System

KW - Medical Imaging

KW - Super Resolution

UR - https://www.scopus.com/pages/publications/105029597596

U2 - 10.1016/j.ifacol.2025.12.552

DO - 10.1016/j.ifacol.2025.12.552

M3 - Conference article

AN - SCOPUS:105029597596

SN - 2405-8971

VL - 59

SP - 649

EP - 654

JO - IFAC-PapersOnLine

JF - IFAC-PapersOnLine

IS - 35

Y2 - 18 November 2025 through 21 November 2025

ER -

DynaFreq-SR: A Dynamic Spatially-Adaptive Network with Frequency-Domain Reconstruction for Chest X-ray Super-Resolution

Abstract

Keywords

Access to Document

Other files and links

Cite this