DCAF-GAN: Enhancing historical landscape restoration with dual-branch feature extraction and attention fusion
PLOS One · 29.10.2025 · 20 (10) · e0334532
DOI: 10.1371/journal.pone.0334532 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS One, online erschienen am 29.10.2025. Verfasst von Li Fang und 4 weiteren Autor:innen. Beteiligte Einrichtung: College of Architecture and Civil Engineering, Qiqihar University, Qiqihar, China. Hauptthemen laut MEDLINE: Deep Learning, Image Processing, Computer-Assisted. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Study on the Construction of Regional Rural Landscape System in Heilongjiang Province, Research on the Path of Empowering the Development of Rural Cultural and Tourism Industry in Heilongjiang Province with New Quality Productivity, Research on the Teaching Effectiveness and Evaluation System Construction of Computer Science and Technology (Higher Vocational and Undergraduate Institutions)….
Abstract (Originalsprache)
Historical landscape restoration has become a crucial area of research in cultural heritage preservation, and with the advancement of digital technologies, effectively restoring damaged historical images has become a critical challenge. Traditional restoration methods face difficulties in handling large occlusions, complex structural features, and maintaining high fidelity in restored images. Existing deep learning methods often focus on restoring a single feature, making it difficult to achieve high-quality reconstruction of both texture and structure. To address these challenges, we propose DCAF-GAN, a novel deep learning model that effectively restores both fine textures and global structures in damaged historical landscapes through a dual-branch encoder and a channel attention-guided fusion module. Experimental results show that DCAF-GAN achieves a PSNR of 29.12 and SSIM of 0.867 on the StreetView dataset, and a PSNR of 28.6 and SSIM of 0.854 on the Places2 dataset, significantly outperforming other models. These results demonstrate that DCAF-GAN not only provides high-quality restorations but also maintains computational efficiency. DCAF-GAN offers a promising solution for the digital preservation and restoration of cultural heritage, with significant potential for further applications.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- College of Architecture and Civil Engineering, Qiqihar University, Qiqihar, China.
- College of Computer and Control Engineering, Qiqihar University, Qiqihar, China.
ORCID-Profile: Li Fang
Förderung
- Study on the Construction of Regional Rural Landscape System in Heilongjiang Province
- Research on the Path of Empowering the Development of Rural Cultural and Tourism Industry in Heilongjiang Province with New Quality Productivity
- Research on the Teaching Effectiveness and Evaluation System Construction of Computer Science and Technology (Higher Vocational and Undergraduate Institutions)…
- Research on CDIO Model for Cultivating Applied Talents in Urban and Rural Planning Major under the New Engineering Education Concept
- Research on the Integration of Party Building and Professional business to Support the Inheritance and Education of Red Genes
Zitieren
APA
Fang, L., Han, B., Bi, M., Wang, L., & Wang, D. (2025). DCAF-GAN: Enhancing historical landscape restoration with dual-branch feature extraction and attention fusion. PLOS One, 20(10), e0334532. https://doi.org/10.1371/journal.pone.0334532
Vancouver
Fang L, Han B, Bi M, Wang L, Wang D. DCAF-GAN: Enhancing historical landscape restoration with dual-branch feature extraction and attention fusion. PLOS One. 2025;20(10):e0334532. doi:10.1371/journal.pone.0334532
BibTeX
@article{Fang2025,
author = {Fang, Li and Han, Bo and Bi, Mingyan and Wang, Lihui and Wang, Dandan},
title = {DCAF-GAN: Enhancing historical landscape restoration with dual-branch feature extraction and attention fusion},
journal = {PLOS One},
year = {2025},
volume = {20},
number = {10},
pages = {e0334532},
doi = {10.1371/journal.pone.0334532}
}
EUVIMED