Ada-DF++: A Dual-Branch Adaptive Facial Expression Recognition Method Integrating Global-Aware Spatial Attention and Squeeze-and-Excitation Attention
Sensors · 24.08.2025 · 25 (17) · 5258
DOI: 10.3390/s25175258 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 24.08.2025. Verfasst von Zhi-Rui Li und 6 weiteren Autor:innen. Beteiligte Einrichtung: School of Information Science and Technology, Hainan Normal University, Haikou 571127, China. Hauptthemen laut MEDLINE: Facial Expression, Attention, Automated Facial Recognition. Die Arbeit wurde bisher 3-mal zitiert (Crossref) und nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Natural Science Foundation of Hainan Province, Major Science and Technology Project of Haikou, Hainan Province Key R&D Program Project.
Abstract (Originalsprache)
Facial Expression Recognition (FER) is a research topic of great practical significance. However, existing FER methods still face numerous challenges, particularly in the interaction between spatial and global information, the distinction of subtle expression features, and the attention to key facial regions. This paper proposes a lightweight Global-Aware Spatial (GAS) Attention module, designed to improve the accuracy and robustness of FER. This module extracts global semantic information from the image via global average pooling and fuses it with local spatial features extracted by convolution, guiding the model to focus on regions highly relevant to facial expressions (such as the mouth and eyes). This effectively suppresses background noise and enhances the model’s ability to perceive subtle expression variations. In addition, we further introduce a Squeeze-and-Excitation (SE) Attention module into the dual-branch architecture to adaptively adjust the channel-wise weights of features, emphasizing critical region information and enhancing the model’s discriminative capacity. Based on these improvements, we develop the Ada-DF++ network model. Experimental results show that the improved model achieves test accuracies of 89.21%, 66.14%, and 63.75% on the RAF-DB, AffectNet (7cls), and AffectNet (8cls) datasets, respectively, outperforming current state-of-the-art methods across multiple benchmarks and demonstrating the effectiveness of the proposed approach for FER tasks.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Information Science and Technology, Hainan Normal University, Haikou 571127, China.
- Hainan eUnited-Tech Technology Co., Ltd., Haikou 570209, China.
Förderung
- Natural Science Foundation of Hainan Province
- Major Science and Technology Project of Haikou
- Hainan Province Key R&D Program Project
- Hainan Province Higher Education Teaching Reform Research Funding Project
- Hainan Province Higher Education Scientific Research Project
- Hainan Province Key R&D Program Project
Zitieren
APA
Li, Z. R., Deng, Z. J., Li, X. Y., Ke, W. D., Yan, S. J., Zhang, J. D., & Liu, C. (2025). Ada-DF++: A Dual-Branch Adaptive Facial Expression Recognition Method Integrating Global-Aware Spatial Attention and Squeeze-and-Excitation Attention. Sensors, 25(17), 5258. https://doi.org/10.3390/s25175258
Vancouver
Li ZR, Deng ZJ, Li XY, Ke WD, Yan SJ, Zhang JD, et al. Ada-DF++: A Dual-Branch Adaptive Facial Expression Recognition Method Integrating Global-Aware Spatial Attention and Squeeze-and-Excitation Attention. Sensors. 2025;25(17):5258. doi:10.3390/s25175258
BibTeX
@article{Li2025,
author = {Li, Zhi-Rui and Deng, Zheng-Jie and Li, Xi-Yan and Ke, Wei-Dong and Yan, Si-Jian and Zhang, Jun-Du and Liu, Chang},
title = {Ada-DF++: A Dual-Branch Adaptive Facial Expression Recognition Method Integrating Global-Aware Spatial Attention and Squeeze-and-Excitation Attention},
journal = {Sensors},
year = {2025},
volume = {25},
number = {17},
pages = {5258},
doi = {10.3390/s25175258}
}
EUVIMED