Geometrical Approaches for Facial Expression Recognition Using Support Vector Machines
Jovan de Andrade Fernandes, Leonardo Nogueira Matos, Maria Géssica dos Santos Aragão · 2016
This article presents two facial geometric-based approaches for facial expression recognition using support vector machines. The first method performed an experimental research to identify the relevant geometric features for human point of view and achieved 85% of recognition rate. The second experiment employed the Correlation Feature Selection and achieved 96.11% of recognition rate. All experiments were carried out with Cohn-Kanade database and the results obtained are compatible with the state-of-the-art in this in this research area.