LOGO-Former: Local-Global Spatio-Temporal Transformer for Dynamic Facial Expression Recognition

• Main Authors: Ma, Fuyan, Sun, Bin, Li, Shutao
• Format: Journal Article
• Language: English
• Published: 05.05.2023
• Subjects: Computer Science - Computer Vision and Pattern Recognition; Computer Science - Multimedia

Previous methods for dynamic facial expression recognition (DFER) in the wild are mainly based on Convolutional Neural Networks (CNNs), whose local operations ignore the long-range dependencies in videos. Transformer-based methods for DFER can achieve better performances but result in higher FLOPs a...