Abstract:

These days, deep learning techniques have seen a lot of success in many different areas, computer vision being just one of them. Convolutional neural network (CNN) models can be trained to perform such analysis and detect emotions in human faces. Using facial features, we train a system to identify how a student is feeling. All together, our system consists of three stages: face detection with Haar Cascades, normalisation, and emotion recognition with CNN on the FER 2013 database with seven different expressions. The results obtained demonstrate the viability of face emotion recognition in the classroom, providing teachers with the means to tailor their presentation to the moods of their students

Description:

CNN, FER2013, Haar Cascade