With the rapid development of computer vision and artificial intelligence, facial expression recognition system has been determinedly investigated, and facial expression recognition system is designed to encode expression information from facial representations.
Automatic facial expression analysis is applied in various practical scenarios such as social robotics, medical treatment, driver fatigue surveillance, and another human-computer interaction system. Recently, deep learning has been regarded as a very expeditious research area and has a high potential to apply in various domains.
Deep learning models are increasingly employed in automatic facial expression recognition to learn discriminative representations. Convolutional Neural Networks (CNN), Deep Belief Networks (DBN), Deep Autoencoder (DAE), Recurrent Neural Networks (RNN), and Generative Adversarial Networks (GAN) are the deep learning models utilized for facial expression recognition.
Deep neural networks are more effectively applied for static and dynamic images in facial expression recognition. Occlusion and non-frontal head pose, infrared data, 3D static and dynamic data, facial expression synthesis, and visualization techniques are some issues in deep learning-based facial expression recognition.
Various literature surveys and reviews on facial expression recognition have been published, presenting deep learning methods, preprocessing approaches, evaluation metrics, benchmark datasets, challenges, opportunities, and future scopes.