Deep Facial Expression Recognition

Abstract

The vast applications of artificial intelligence, such as human-computer collaboration, data-driven animation, human-robot interaction, etc., have made it urgently necessary to detect emotions through facial expression. Numerous studies have been done on this subject since it is a challenging and intriguing issue in computer vision. The goal of this study is to create a deep convolutional neural network-based face expression recognition system with data augmentation. This method makes it possible to identify the seven fundamental emotions which are anger, disgust, fear, happiness, neutrality, sadness, and surprise. Recent years have seen a rise in interest in the field of facial recognition technology, and convolutional neural networks (CNNs), which have demonstrated outstanding achievements in this domain. CNN training may take a long time and it requires a lot of labeled data. Nonetheless, in this paper, we have investigated the application of transfer learning methods to enhance the effectiveness and precision of CNN-based facial recognition tasks. Besides, We have looked at how well feature extraction and fine-tuning pre-trained models work together to transfer knowledge from one domain to another.