| In recent years,the integration of the Internet and education has come into our vision.It has become a rising star in the education industry.This integration mainly focuses on online video learning.Depend on the integration of lesson resources,people could learn the knowledge they are interested by online video classrooms at home.Students could learn more independently.It enriches the new teaching mode for the education industry.At present,only the teacher can impart knowledge to the students on online video teaching platform.The listening status of students can’t be observed by teachers in time.This kind of supervision mechanism is deficient in the teaching process.In order to bring students better classroom experience,online video teaching platforms should detect and evaluate students’ classroom status.Feedback the results to teachers in order to adjust the progress and difficulty of the course appropriately.At this time,it is very important to detect the learning status of students in online classrooms,and the most important part is fatigue detection.In this thesis,a face fatigue detection model is proposed.This model is applied to online children’s programming teaching class.The main function of the model is to detect the face and fatigue of the students in the online course.It could judge whether the person in the video is tired and return the result.The model is used in the online teaching platform.Fatigue detection of students in class is implemented by the communication between the server and the client.After investigating a variety of face detection and video classification methods in order to finish the face fatigue detection model,finally deciding to adopt a deep learning-based method.Selecting Face R-CNN for face detection and using dual-stream convolutional neural network as video classification Framework.Combining the two methods and using public data sets to train the model and evaluate and optimize the performance of the model.After obtaining the trained model,this article combines the model with an online teaching platform based on Angular2 to form a student fatigue detection system,and evaluates the system,and proposes an optimization system plan from both real-time and accuracy.Finally,in the laboratory environment,it was verified through the video recording data of the online classroom which showed good fatigue detection effect. |
