Online Group-Based Face Recognition and Tracking Algorithm Using Kalman Filtering in Team Learning Mode

Since the 21st century, with the advancement of the times, the educational model has changed, and a new teaching model, the team learning model, has emerged. However, traditional methods of educational assessment are inadequate to meet the new demands. Obtaining personalized characteristics of learners in the classroom is essential for achieving precision, objectivity, personalization, and intelligent education. To efficiently collect students' facial images, this paper proposes an online group face recognition and tracking algorithm based on Kalman filtering. The algorithm considers the continuity of facial targets in time and the correlation in space, addressing the impact of partial facial poses, occlusions, and expressions on face detection and recognition. Moreover, by utilizing the predictive function of the Kalman filter, the algorithm reduces the number of face feature comparisons, improving the timeliness and accuracy of face recognition. Experimental results validate the effectiveness of this algorithm.