融合VGG与注意力的学生微表情识别和情绪评估方法  

Student micro expression recognition and emotion assessment based on the integrating of attention and VGG

作　　者：刘芳[1] 李俊吉[1] Liu Fang;Li Junji(College of Information Science and Technology,Taiyuan University of Science and Technology,Jincheng 048011,China)

机构地区：[1]太原科技大学信息科学与技术学院,晋城048011

出　　处：《现代计算机》2024年第18期28-33,共6页Modern Computer

基　　金：2023年山西省高等学校一般性教学改革创新项目(J20230880)。

摘　　要：在智能课堂中,实时掌握学生的情绪状态对于提高教学质量和个性化教育具有重要意义。引入通道注意力机制,对VGG16卷积神经网络进行改进,结合多层感知机,提出了VGG16_SE_MLP模型用于学生微表情分类识别以及情绪评估方法。首先对微表情数据集进行预处理,然后进行特征提取,在卷积层后面引入SE模块,并加入批归一化层防止过拟合,通过MLP计算得到新的特征向量以及微表情类别,最后对学生情绪进行评估。实验结果表明,该方法在微表情分类识别和情绪评估效果性能良好,为智能课堂提供了新思路。In the intelligent classroom,it is of great significance to grasp the emotional state of student in real time for improving the quality of teaching and personalized education.The channel attention mechanism is introduced to improve the VGG16 convolutional neural network.Combined with the multilayer perceptron the VGG16_SE_MLP model is proposed for the classification and recognition of student micro‑expression and emotion assessment.Firstly,the micro‑expression dataset is preprocessed,then the feature is extracted,and the SE module is introduced behind the convolution layer,and the batch normalization layer is added to prevent overfitting;the extracted features and micro‑expression class are calculated layer by layer through MLP;finally,the student emotion is evaluated.Experimental results show that the proposed method performs well in micro‑expression classification and emotional assessment,providing a new idea for intelligent classroom.

关 键 词：微表情识别 通道注意力机制 VGG16卷积神经网络 多层感知机 批归一化层 情绪评估 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] G434[自动化与计算机技术—计算机科学与技术]