Diffusion-based Real-World Super-Resolution
Adapting pretrained diffusion models to real-world super-resolution with degradation-aware representations and one-step restoration.
View project →Research
Research & publications
Image restoration, low-light enhancement, super-resolution and object detection — from lightweight models for edge devices to diffusion-based restoration.
8Projects
14Publications & preprints
3Award-winning projects
Currently
Diffusion models for real-world super-resolution at NYCU's ACM Lab.
Projects
Dynamic Illuminance Enhancement for Diverse Scenes
Developing smart systems to handle extreme lighting variability in real-world environments.
View project →
Lightweight Deep Learning for Image Enhancement
Highly efficient neural networks designed for mobile devices and edge computing.
View project →
Weather Image Generation (Thesis)
End-to-End deep learning network using Triplet Content Loss for weather synthesis.
View project →
Image Dehazing using SR Networks
Joint training of dehazing and super-resolution models to reduce computation.
View project →
Hairstyle Recommendation System
Face Shape Classification and Swapping for personalized recommendation.
View project →
Drone-based Aerial Image Super-Resolution
Improving the resolution of drone-based aerial images for better object detection.
View project →
Mask Detection & People Flow Systems
Real-time public safety monitoring with YOLO-based detection and tracking.
View project →Publications
2026
Bridging Restoration and Generation in One-step Diffusion for Real-World Image Super-Resolution
arXiv preprint · Apr 2026
The First Challenge on Mobile Real-World Image Super-Resolution at NTIRE 2026: Benchmark Results and Method Overview
arXiv preprint · Apr 2026
2025
Learning Ordinal Degradation Representations with Textual Priors for Diffusion-Based Blind Image Super-Resolution
arXiv preprint · Dec 2025
A lightweight low-light image enhancement network via channel prior and gamma correction
International Journal of Pattern Recognition and Artificial Intelligence (IJPRAI), vol. 39, no. 12 · 2025
An Integrated Approach to Mask Wearing Classification and Crowd Counting in Public Spaces Using YOLOv5s and Deep SORT
Journal of Internet Technology, vol. 26, no. 4, pp. 423-434 · Jul 2025
2024
Rethinking Theoretical Illumination for Efficient Low-Light Image Enhancement
arXiv preprint · Sep 2024
Dynamic Illuminance Adjustment for Simultaneous Exposure Correction and Low-Light Image Enhancement
Signal, Image and Video Processing · Sep 2024
Scene-Guided Image Enhancement through Gamma Correction Prior and Transfer Learning for Varied Illumination Environments
CVGIP 2024, Hualien, Taiwan · Aug 2024
Exposure Correction in Driving Scenes Using the Atmospheric Scattering Model
IEEE ICCE-TW 2024, Taichung, Taiwan · 2024
Weather Image Generation (DGNDW)
Master's Thesis, CYCU · 2024
2023
Reducing Computational Requirements of Image Dehazing Using Super-Resolution Networks
IEEE IS3C 2023, Taichung, Taiwan · 2023
Using Stereo Depth Estimation Network and LiDAR-Assisted Camera for Dehazing
2023 9th IEEE International Conference on Applied System Innovation (ICASI2023), Chiba, Japan · 2023
Improving Hair Design Efficiency and Customer Satisfaction: A Hairstyle Recommendation System with Advanced Face Analysis, Face Swapping, and Hair Dyeing Techniques
CVGIP 2023, Taiwan · Aug 2023
A Non-Adversarial and End-to-End Deep Learning Generative Network for Generating Diverse Weather Images
SSRN preprint · Jan 2023
Peer review ORCID
IEEE Transactions on Image ProcessingIEEE Transactions on Circuits and Systems for Video TechnologyDigital Signal ProcessingScientific ReportsDiscover ElectronicsJournal of King Saud University - Computer and Information SciencesThe Imaging Science Journal