Deep Learning-Based Facial Expression Recognition

Abstract

Facial Expression Recognition is a deal in computer vision and emotional intelli-gence. It helps computers understand how humans feel and interact with them in a natural way. Traditional methods have trouble working outside of labs but deep learning can help computers understand human emotions from videos and pictures. This research is about creating Facial Expression Recognition systems using the latest deep learning techniques and transfer learning. It aims to make computers better, at understanding emotions. This thesis is a study of Facial Expression Recognition using Deep Learning and Transfer Learning. It starts by explaining the basics of Facial Expression Recognition, including how emotion-sre represented, how machines process facial expressions and how traditional machine learning methods work. Then it looks at learning, including neural networks and convolutional neural net-works. It also looks at transfer learning, which’s a way to adapt pre-trained models to new jobs. Several pre-trained models, including AlexNet, VGG and GoogleNet are tested to see how well they can recognize expressions. The results show that transfer learning is really good, at making recognition while making training easier and faster. Finally The findings suggest that deep learning can be used to make Facial Expression Recognition systems that work in real life