An intelligent teaching assistant system using deep learning technologies
Zheyu Zhou · 2020
In this paper, we describe an intelligent teaching assistant system using deep learning technologies. A few works had been done on building intelligent assistant for teachers before and our job is novel. The main challenge in this area is the distraction detection under various surroundings of the students. We invent a creative way to embed human prior knowledge of distraction judgment. Our system takes multiple sequential images as input, runs through a prior feature image extraction component using image segmentation and face detection technologies base on deep learning, and then a distraction detection component which uses AlexNet to do the classification, and finally outputs the evaluation of the online lesson. Our system has achieved 85.8% precision on student distraction detection and the evaluation generated by the system can serve as an indication to notify the teacher when specific teaching methods should be taken so as to enhance the lesson effectiveness.