Scholarlink

Facial Expression Recognition (FER) is pivotal in identifying human psychological states across various domains. In educational settings, recognizing students’ emotions in real time can support personalized teaching, improve learning outcomes, and increase student engagement. Despite recent advancements in FER, limited focus has been given to hybrid models incorporating data augmentation techniques to enhance robustness. In this research, we introduce Extended EfficientFormer (ExtEF), a lightweight hybrid method incorporating data augmentation, specifically designed for facial expression recognition (FER) in personalized learning settings. This method combines the efficiency of the base EfficientFormer with enhanced feature representation through additional residual blocks. Comprehensive experiments on the FER-2013 dataset demonstrate that the proposed approach improves performance. A comparison with the most popular methods is performed, showcasing superior accuracy, robustness, and computational efficiency. These findings highlight the suitability of the proposed approach for FER tasks in educational contexts, enabling real-time emotion monitoring to personalize learning experiences and foster student engagement in adaptive learning systems and virtual classrooms.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing Facial Expression Recognition for Personalized Learning: A Hybrid Approach with Data Augmentation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Youssef Amkrane,
  • Mohamed Badiy,
  • Fatima Amounas,
  • Mourade Azrour,
  • Salma Bendaoud

摘要

Facial Expression Recognition (FER) is pivotal in identifying human psychological states across various domains. In educational settings, recognizing students’ emotions in real time can support personalized teaching, improve learning outcomes, and increase student engagement. Despite recent advancements in FER, limited focus has been given to hybrid models incorporating data augmentation techniques to enhance robustness. In this research, we introduce Extended EfficientFormer (ExtEF), a lightweight hybrid method incorporating data augmentation, specifically designed for facial expression recognition (FER) in personalized learning settings. This method combines the efficiency of the base EfficientFormer with enhanced feature representation through additional residual blocks. Comprehensive experiments on the FER-2013 dataset demonstrate that the proposed approach improves performance. A comparison with the most popular methods is performed, showcasing superior accuracy, robustness, and computational efficiency. These findings highlight the suitability of the proposed approach for FER tasks in educational contexts, enabling real-time emotion monitoring to personalize learning experiences and foster student engagement in adaptive learning systems and virtual classrooms.