基于面部表情和头部姿态的学习者情绪分析与评价研究

A Study of Learners’ Emotion Analysis and Evaluation Based on Facial Expressions and Head Posture

利用计算机视觉技术分析和评价学习者情绪是当前智能教育领域的重点研究的方向。由于课堂人数众多,仅凭教师和督导的观察与评价,无法全面、实时掌握每个学生的学习情况。因此,本文分别结合面部表情和头部姿态检测结果,并结合真实课堂的特点,对学习者情绪进行分类,然后加权融合,智能化地将学生的课堂实时情绪分为积极、中性和消极3个类别,又进一步以学时为单位将学生的情绪细分为7个类别。通过设计学习者情绪分析与评价系统,并在真实智慧课堂场景中使用测试,分析和评价数据表明,该模型能真实反映学生的情绪,为教师深化教学改革等提供客观有效的依据。

Using computer vision technology to analyze and evaluate learners’ emotions is a key research direction in the field of intelligent education, realizing the transformation and upgrading of traditional classrooms empowered by data. Due to the large number of students in the classroom, only the observation and evaluation of teachers and supervisors cannot comprehensively and real-time grasp the learning situation of each student. Therefore, this paper combines the detection results of facial expressions and head poses,and combines the characteristics of real classrooms to classify learners’ emotions, and then weights and fuses them to intelligently divide students’ real-time classroom emotions into positive, neutral, and negative, and further subdivided students’ emotions into 7categories by credit hours. By designing a learner emotion analysis and evaluation system, and using tests in real smart classroom scenarios, analysis and evaluation data show that the model can truly reflect students’ emotions and provide an objective and effective basis for teachers to deepen teaching reform.