Student Action Recognition and Early Warning Machine Based on Online Class
Xuefei Lv, Wenhui Zhang · 2021 IEEE 3rd International Conference on Frontiers Technology of Information and Computer (ICFTIC) · 2021
In order to ensure the students' listening status, this paper proposes and implements a student action recognition and early warning mechanism based on online class. Firstly, use Kinect to collect the time series data of the human body action feature description, and construct the human body posture description vector; Then, use the principal component analysis (PCA) method to extract different action features, reduce the dimension of the feature vector space, eliminate the difference between the original features, and reduce the redundancy of data information in order to judge and recognize actions; Finally, for the identified actions that are not conducive to listening to the class, the warning tone is used to remind and wake up the students who are not paying attention in class.