USK-FEMO: A Face Emotion Dataset using Deep Learning for Effective Learning

Emotions play an essential role in the learning process and have an impact on how the learning process is eventually carried out. Facial expressions can be used to visually identify a person’s emotions. Along with the advancement of computer vision and deep learning techniques, the study of human-computer interaction is increasingly focusing on the recognition of facial expressions. One of the main issues is the availability of sufficient datasets, especially for students. This study examined the deep learning architecture for face emotion classification. In addition, this research also introduces a new emotional dataset acquired from the junior high school student at SMP Negeri 1 Darul Imarah, Aceh Besar Regency, Indonesia. This dataset contains five classes that include the emotions of happiness, sadness, anger, surprise, and boredom. The dataset was then tested using the Mobile-Net architecture, the highest accuracy was achieved with a learning rate of 0.0001% of 88.492%. The dataset can be explored via the link https://muhajir2111.github.io/USK-FEMO-DATASET/