Facial recognition of emotions in higher level students during remote classes (#1570)

Authors

Tiznado Ubillus, Jose Armando

Rivera Escriba, Luis Antonio

Huamán Martínez, Milagros Florencia Mercedes

Velásquez Oyola, Margarita Betzabé

Salazar Fierro, Fausto Alberto

Arboleda Huaman, Julio Fernando

Atoche Pacherres, César Augusto

Abstract

During remote classes, teachers face the challenge of recognizing students' emotions, especially when their cameras are on. To address this problem, a facial detection and recognition system was developed that allowed measuring emotions in real time, providing data that can be statistically analyzed. The objective was to establish and classify the emotions expressed by students during remote classes, which allows detecting their real-time participation in the virtual environment. To obtain the best predictive model for detecting student participation in real time, different models were adjusted. Various reference data sets were used to measure the performance and accuracy of the proposed system. The results showed that the proposed system achieves accuracy on different reference sets and on the proposed data set itself. The real-time predictive classification model, modelFEC6.h5, outperforms the others with an accuracy of 94.62% for facial emotion classification in real-time virtual learning scenarios.