Automatic Academic Confusion Recognition In Online Learning Based On Facial Expressions

Academic confusion is one of the most common academic emotions, timely identification and resolution of confusion are helpful to improve learning effect. In the research, a model is developed to identify the academic confusion in online learning based on facial expressions. The model mainly includes three parts: confusion-inducing experiments, image preprocessing, and recognition methods comparing. Firstly, a set of confusion-inducing experiments in online learning are designed. Then the images are preprocessed to improve the recognition effect. In the third part, Histogram of Oriented Gradient (HOG), Local Binary Patterns (LBP), Support Vector Machine (SVM) and Convolutional Neural Network (CNN) are combined to form four methods of HOG-SVM, LBP-SVM, CNN and CNN-SVM, which are employed. The experimental results show that most of the methods can effectively detect students' academic confusion, and the CNN-SVM has the best predictive performance with an average accuracy of 0.938. The Model proposed in this study can provide a technical support for learner emotional modeling of teaching assistant systems.