Edwyn Martínez, Edmundo Bonilla, Eduardo Sánchez, Crispín Hernández and Roberto Morales, Tecnológico Nacional de Mexico, Apizaco, Mexico
This paper proposes to improve the Viola-Jones algorithm to identify facial features in thermal images using a fuzzy linguistic modifier. The purpose is to improve the image quality in the infrared plane and to facilitate the recognition stage, feature selection and emotion classification. The results obtained show that adding a preprocessing stage improves the capability of the Viola-Jones algorithm in identifying facial features in thermal images.
Emotion recognition, Edge detection, Viola-Jones, Image processing.