M3ENet: A Multi-Modal Fusion Network for Efficient Micro-Expression Recognition
Sensors · 10.10.2025 · 25 (20) · 6276
DOI: 10.3390/s25206276 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, online erschienen am 10.10.2025. Verfasst von Ke Zhao und 2 weiteren Autor:innen. Beteiligte Einrichtung: Electronic Science and Technology Museum, University of Electronic Science and Technology of China, Chengdu 611731, China. Hauptthemen laut MEDLINE: Facial Expression, Pattern Recognition, Automated. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 53 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Ministry of Education of China Industry, Ministry of Education of China Industry-University Cooperative Education Program, Sichuan Science and Technology Program.
Abstract (Originalsprache)
Micro-expression recognition (MER) aims to detect brief and subtle facial movements that reveal suppressed emotions, discerning authentic emotional responses in scenarios such as visitor experience analysis in museum settings. However, it remains a highly challenging task due to the fleeting duration, low intensity, and limited availability of annotated data. Most existing approaches rely solely on either appearance or motion cues, thereby restricting their ability to capture expressive information fully. To overcome these limitations, we propose a lightweight multi-modal fusion network, termed M3ENet, which integrates both motion and appearance cues through early-stage feature fusion. Specifically, our model extracts horizontal, vertical, and strain-based optical flow between the onset and apex frames, alongside RGB images from the onset, apex, and offset frames. These inputs are processed by two modality-specific subnetworks, whose features are fused to exploit complementary information for robust classification. To improve generalization in low data regimes, we employ targeted data augmentation and adopt focal loss to mitigate class imbalance. Extensive experiments on five benchmark datasets, including CASME I, CASME II, CAS(ME)2, SAMM, and MMEW, demonstrate that M3ENet achieves state-of-the-art performance with high efficiency. Ablation studies and Grad-CAM visualizations further confirm the effectiveness and interpretability of the proposed architecture.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Electronic Science and Technology Museum, University of Electronic Science and Technology of China, Chengdu 611731, China.
- School of Public Administration, University of Electronic Science and Technology of China, Chengdu 611731, China.
- State Key Laboratory of Photonics and Communications, Peking University, Beijing 100871, China.
- Department of Biomedical Engineering, Hong Kong Polytechnic University, Hong Kong 999077, China.
ORCID-Profile: Ke Zhao, Xuanyu Liu, Guangqian Yang
Förderung
- Ministry of Education of China Industry
- Ministry of Education of China Industry-University Cooperative Education Program
- Sichuan Science and Technology Program
- AI Technology-Enabled Undergraduate Teaching Reform Projects at the University of Electronic Science and Technology of China
- Sichuan Provincial Science Popularization Base Assessment Project
- Graduate Education and Teaching Reform Program at UESTC
Zitieren
APA
Zhao, K., Liu, X., & Yang, G. (2025). M3ENet: A Multi-Modal Fusion Network for Efficient Micro-Expression Recognition. Sensors, 25(20), 6276. https://doi.org/10.3390/s25206276
Vancouver
Zhao K, Liu X, Yang G. M3ENet: A Multi-Modal Fusion Network for Efficient Micro-Expression Recognition. Sensors. 2025;25(20):6276. doi:10.3390/s25206276
BibTeX
@article{Zhao2025,
author = {Zhao, Ke and Liu, Xuanyu and Yang, Guangqian},
title = {M3ENet: A Multi-Modal Fusion Network for Efficient Micro-Expression Recognition},
journal = {Sensors},
year = {2025},
volume = {25},
number = {20},
pages = {6276},
doi = {10.3390/s25206276}
}
EUVIMED