Title: VIVAFER - voluntary and involuntary actions-based facial expression and recognition

Authors: Annadurai Swaminathan; Michael Arock

Addresses: Department of Computer Applications, National Institute of Technology, Tiruchirappalli, Tamil Nadu, 620015, India ' Department of Computer Applications, National Institute of Technology, Tiruchirappalli, Tamil Nadu, 620015, India

Abstract: Facial expression recognition (FER) and analysis play a vital role in developing emotion and FER-based applications. In past research, anger, surprise, contempt, happy, disgust, sad, fear and neutral were the basic emotions usually inferred from FER. Humans, also exhibit some facial expressions based on voluntary and involuntary actions (VIVA), which do not infer basic emotions. VIVAFER must be considered as essential expressions when developing an application. Otherwise, if a yawn (a class in VIVAFER) is considered as 'surprised', then the application becomes meaningless. Very few researchers dealt with this area, which is also limited to specific applications. Considering the above problem, the present work broadly classifies FER based on VIVA. For research purposes, a new dataset named VIVAFER is constructed. Various machine and deep learning algorithms are used for training, and testing of these 15 classes. The applications of VIVAFER are behavioural analysis, feedback system, medical-applications, patient monitoring system, etc.

Keywords: action-based expression; facial expression; facial expression recognition; FER; facial expression analysis; involuntary action-based facial expression; voluntary action-based facial expression.

DOI: 10.1504/IJBIR.2021.119815

International Journal of Business Innovation and Research, 2021 Vol.26 No.4, pp.467 - 487

Received: 09 Sep 2019
Accepted: 01 Dec 2019

Published online: 21 Dec 2021 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article