AI-Based Facial Emotion Recognition

One of the most ubiquitous ways that living beings communicate is through facial emotions. On that account emotion recognition is crucial for the interaction between computer and humans. Due to the complexity of human faces and differences in photographs, such as varied facial poses and lighting, facial emotion recognition by computer models still poses a challenge. Owing to their effective computational performance and excellent autonomous extraction of features, deep learning models, particularly Convolutional Neural Networks (CNNs), have demonstrated considerable potential among all facial emotion recognition techniques. In this paper, we have proposed an improved version of the Visual Geometry Group-16 (VGG-16) model on the very challenging Facial emotion recognition (FER) 2013 dataset. The proposed model achieved validation accuracy of 85.92%.