Automated Facial Expression Recognition (FER) is an important part of computer-human interaction. For decades, researchers and scientists have been trying to create a model of artificial intelligence that could think, learn, make decisions and act in a way similar to a real person. Among other skills, such model needs to recognise human facial expression to understand non-verbal language. The present paper describes a method to fine tune the FER process in images, using deep learning CNN model Xception, with preprocessing the images. The method has shown improved results when applied to different datasets.
Digital Object Identifier (DOI)
Kanatov, Maksat; Atymtayeva, Lyazzat; and Mendes, Mateus
"Improved Facial Expression Recognition with Xception Deep Net and Preprocessed Images,"
Applied Mathematics & Information Sciences: Vol. 13:
5, Article 20.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol13/iss5/20