Facial expression recognition by using differential geometric features

ABSTRACT

In recent years, a growing interest has been created for improvement of human interaction with computers. Hence, automatic recognition of facial expressions has become one of the active research topics. The purpose of this paper is to identify facial expressions, by using differential geometric features. In the proposed method, only the first and last images are used and differential features are extracted from these two images. Differential geometric features are extracted from changes in the important points of the face in the two images. In this method, the distance between the important points of the face and the reference point was calculated in both directions x and y, for two images, and with the difference between the distance, the differential geometric features between the two images were obtained. Based on the results, with this method, recognition accuracy of six facial expressions in the database was 96.44%, CK +.

KEYWORDS:

• Facial expression
• emotional images
• differential geometric features
• recognition accuracy
• database CK +

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Aref Moradi http://orcid.org/0000-0002-1731-5979

Notes on contributors

Erfan Zangeneh received his B.S. degree in Computer Engineering from Buali Sina, Hamedan, Iran in 2013 and M.S. degrees in Computer Engineering (Artificial Intelligence) from Amirkabir University of Technology, Tehran, Iran in 2015. He now works at Faraadid as a computer vision expert.

Aref Moradi received his B.S. degree in Software Engineering from Amirkabir University of Technology in 2018. He worked one year at CaféBazaar as a Data Scientist. He is currently attending his master studies in Data Science in Engineering at Eindhoven University of Technology.

Image note

The image dataset of figures 1, 2, 3 and 5 can be accessed through the following link:

http://www.consortium.ri.cmu.edu/ckagree/