Students Engagement Detection Using Hybrid EfficientNetB7 Together With TCN, LSTM, and Bi-LSTM (DAiSEE and VRESEE datasets)

Model Development with the DAiSEE Dataset for an Intelligent Tutoring System recognizing Comprehension Problems during Learning

A multi-view user engagement recognition prototype classifying engagement levels of 4. Developed Convolution Neural Network extending VGG-16 with effective weights on DAiSEE dataset with an accuracy of 87%.  Enhanced stability with micro and macro expressions of users from an unstructured e-learning environment.