Research and Implementation of Key Technologies for Real-time Classroom Feedback System Based on Artificial Intelligence in Higher Education

The study created a new real-time feedback system, primarily for college classrooms, to observe how students learn during traditional classes. The system has several layers of structure that make the monitoring of the teaching process smarter by collecting multiple data and deep learning methods. It has several special features: first, an improved, lightweight convolutional neural network, which can better recognize face expressions and can be used in different environments; secondly, a attention evaluation model that combines multiple functions to determine whether students are listening carefully; and finally, optimizing the structure of the entire system.It's 89.7%, and it takes less than 28 milliseconds to process a video stream. After actually eight weeks, it was found that students' participation in the classroom was 21% higher than before, and 90% of teachers felt that the system was good. These data show that the system is really useful.