Effective image compression using transformer and residual network for balanced handling of high and low-frequency information
PLOS One · 2025
DOI: 10.1371/journal.pone.0333376 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS One, 2025. Verfasst von Jianhua Hu und 5 weiteren Autor:innen. Beteiligte Einrichtung: Computer Engineering Technical College, Guangdong Polytechnic of Science and Technology, Zhuhai, Guangdong, China. Hauptthemen laut MEDLINE: Data Compression, Image Processing, Computer-Assisted, Neural Networks, Computer. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 25 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Foundation for 2022 Basic and Applied Basic Research Project of Guangzhou Basic Research Plan, Research and Practice on AI-based Classroom Teaching Evaluation Reform, 2024 Guangdong Provincial Junior Innovative Talents Project for Ordinary Universities.
Abstract (Originalsprache)
Image compression has made significant progress through end-to-end deep-learning approaches in recent years. The Transformer network, coupled with self-attention mechanisms, efficiently captures high-frequency features during image compression. However, the low-frequency information in the image cannot be obtained well through the Transformer network. To address this issue, the paper introduces a novel end-to-end autoencoder architecture for image compression based on the transformer and residual network. This method, called Transformer and Residual Network (TRN), offers a comprehensive solution for efficient image compression, capturing essential image content while effectively reducing data size. The TRN employs a dual network, comprising a self-attention pathway and a residual network, intricately designed as a high-low-frequency mixer. This dual-network can preserve both high and low-frequency features during image compression. The end-to-end training of this model employs rate-distortion optimization (RDO methods). Experimental results demonstrate that the proposed TRN method outperforms the latest deep learning-based image compression methods, achieving an impressive 8.32% BD-rate (bit-rate distortion performance) improvement on the CLIC dataset. In comparison to traditional methods like JPEG, the proposed achieves a remarkable BD-rate improvement of 70.35% on the CLIC dataset.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Computer Engineering Technical College, Guangdong Polytechnic of Science and Technology, Zhuhai, Guangdong, China.
- College of Computer Science, Guangdong Polytechnic Normal University, Guangzhou, Guangdong, China.
ORCID-Profile: Guixiang Luo, Wei Nie
Förderung
- Foundation for 2022 Basic and Applied Basic Research Project of Guangzhou Basic Research Plan
- Research and Practice on AI-based Classroom Teaching Evaluation Reform
- 2024 Guangdong Provincial Junior Innovative Talents Project for Ordinary Universities
- 2023 Guangdong Provincial Junior Innovative Talents Project for Ordinary Universities
- Innovation Team Project of Colleges and Universities of Guangdong Provincial Department of Education
- Research and implementation of image defect recognition algorithm for power line inspection by drone based on Ascend NPU
- Guangdong Intelligent Vocational Education Engineering Technology Research Center
- 2020 Guangdong Provincial Special Innovation Project for Ordinary Universities
Zitieren
APA
Hu, J., Luo, G., Feng, X., Yuan, Z., Yang, J., & Nie, W. (2025). Effective image compression using transformer and residual network for balanced handling of high and low-frequency information. PLOS One. https://doi.org/10.1371/journal.pone.0333376
Vancouver
Hu J, Luo G, Feng X, Yuan Z, Yang J, Nie W. Effective image compression using transformer and residual network for balanced handling of high and low-frequency information. PLOS One. 2025. doi:10.1371/journal.pone.0333376
BibTeX
@article{Hu2025,
author = {Hu, Jianhua and Luo, Guixiang and Feng, Xiangfei and Yuan, Zhanjiang and Yang, Jiahui and Nie, Wei},
title = {Effective image compression using transformer and residual network for balanced handling of high and low-frequency information},
journal = {PLOS One},
year = {2025},
doi = {10.1371/journal.pone.0333376}
}
EUVIMED