Research and Implementation of Student Attention Detection Based on Head Pose Estimation

• 1、School of Computer and Cyber Sciences, Communication University of China, Beijing 100024

Abstract：With the popularization and promotion of online education in China, more and more schools and institutions have carried out online education classes. Unlike "face-to-face" traditional education, online education can not make an intuitive judgment on of students\' attention, it is difficult to understand students\' learningstate and performance in time, the teaching quality can not be guaranteed, therefore, it is of great research value and practical significance to introduce the detection of students\' attention into online education. There are many factors to judge whether the students are paying attention to the class. Among them, the rotation angle of the head can directly reflect the students\' learning state and attention in class,the key technology of head pose estimation is face recognition and key points detection. However, in traditional key points detection methods, the detection results are not accurate due to the external environment factors such as occlusion, light intensity, shooting angle and so on. This paper proposed a method based on multi-task cascaded convolution neural networks and head pose estimation, designed a students\' attention detection system to to detect the students\' attention in real time and feedback the results, finally achieved the purpose of detecting students\' attention in online learning.

Keywords： head pose estimation face key points detection attention detection online learning