| Students’ emotions in the classroom reflect their learning status,and teachers can adjust their teaching methods and improve teaching efficiency through students’ emotional performance.However,due to the large number of students in the classroom,teachers are unable to effectively take into account the perception of students’ learning emotions while teaching;at the same time,the traditional teaching quality evaluation method uses lecture evaluation and homework grades as indicators,ignoring the importance of students’ learning status.Therefore,using artificial intelligence technology to help teachers obtain students’ emotional state in real time and identify students’ learning state is very important for teachers’ teaching activities arrangement and teaching quality evaluation.Facial expressions reflect human emotions and contain rich information,which can reflect human emotional state in real time,so using artificial intelligence technology to recognize students’ facial expressions can realize student learning state monitoring.In this paper,we propose a lightweight expression recognition model based on convolutional attention to achieve face expression recognition,and then propose to convert expressions into PAD values from the perspective of educational psychology,and then use PAD values to evaluate students’ learning status.In order to increase the diversity of learning status determination criteria,head-up rate detection and fatigue status detection are introduced in this paper.The head-up rate detection is based on face recognition technology,and the head-up rate can be obtained by calculating the ratio of the number of faces to the total number of students.Fatigue detection is based on face key point detection,and the frequency and duration of blinking and yawning actions are used to determine whether a person is in a fatigue state.Finally,the expression recognition,head-up rate and fatigue status are weighted together to evaluate the students’ learning status comprehensively.Based on the comprehensive evaluation method of learning status,this paper designs and develops a real-time learning status monitoring system that can run on small terminal devices such as Raspberry Pi and PC,and can detect local video or camera video stream in real time.In order to verify the effectiveness of the system,this paper carries out validation experiments by synthesizing videos with sequential pictures from BNU-LSVED 2.0 teaching quality evaluation dataset.The experimental results show that the learning status detection method based on facial expression recognition proposed in this paper can identify students’ learning status more accurately and has strong practicality. |
