Face Recognition Using Dynamic Features Extracted from Smile Videos

Biometric systems measure and analyze the physical and behavioral characteristics of individuals. Facial biometry has become one of the most preferred biometric methods in recent years due to the fact that it can be used in applications, which do not require the cooperation of the user. In facial recognition systems, which use images recorded under controlled conditions, the use of physical properties is sufficient for face recognition. However, physical features may not be sufficient for face recognition using images recorded under challenging conditions. In such cases, behavioral characteristics of the face may provide additional information. In this study, we extract dynamic features of the face from expressive face videos and use them for face recognition. Experimental results on the UvA-NEMO smile database, which contains 400 subjects, show that dynamic face features carry identity-related information and can be used for face recognition.