EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Real-World Adaptation of Retinexformer for Low-Light Image Enhancement Using Unpaired Data

Subhan Uddin, Babar Hussain, Sidra Fareed, Aqsa Arif, Babar Ali

International Journal of Ethical AI Application · 2025

DOI: 10.64229/6rrx1t06 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Ethical AI Application, 2025. Verfasst von Subhan Uddin und 4 weiteren Autor:innen. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Low-light image enhancement remains a significant challenge in real-world computer vision applications, especially where lighting conditions vary drastically and paired training data is unavailable. While transformer-based models have shown promise in controlled environments, their ef-fectiveness often diminishes when applied to naturally degraded images. This paper presents a novel approach for adapting a transformer-based enhancement model to realworld low-light scenarios using unpaired datasets. We utilize real low-light images captured under uncontrolled conditions and propose a domain adaptation framework that enables effective transfer learning from synthetic to real domains. Our method integrates unsupervised reconstruction loss, perceptual optimization, and domain-invariant feature alignment to refine the model’s performance without requiring paired supervision. Experimental evaluations reveal notable improvements in both visual quality and quantitative metrics on real-world benchmarks. Compared to existing enhancement methods, our approach offers superior generalization, robustness to noise, and high-fidelity out-put. This demonstrates the potential of our domain-adapted transformer model in practical low-light imaging applications, including night photography, surveillance, and mobile vision systems.

Zitieren

APA

Uddin, S., Hussain, B., Fareed, S., Arif, A., & Ali, B. (2025). Real-World Adaptation of Retinexformer for Low-Light Image Enhancement Using Unpaired Data. International Journal of Ethical AI Application. https://doi.org/10.64229/6rrx1t06

Vancouver

Uddin S, Hussain B, Fareed S, Arif A, Ali B. Real-World Adaptation of Retinexformer for Low-Light Image Enhancement Using Unpaired Data. International Journal of Ethical AI Application. 2025. doi:10.64229/6rrx1t06

BibTeX
@article{Uddin2025,
  author = {Uddin, Subhan and Hussain, Babar and Fareed, Sidra and Arif, Aqsa and Ali, Babar},
  title = {Real-World Adaptation of Retinexformer for Low-Light Image Enhancement Using Unpaired Data},
  journal = {International Journal of Ethical AI Application},
  year = {2025},
  doi = {10.64229/6rrx1t06}
}
Kostenlos starten →