A Key Frame Selection-Based Facial Expression Recognition System

Automatic recognition of facial expressions has received considerable interest in the computer vision. This paper presents a key frame selection-based system for recognition of four basic facial expressions: happiness, anger, sadness and surprise, in image sequences. The proposed new approach is designed for recognition of facial expressions via a key frame selection of the facial motion from an image sequence where an optical flow is used to determine changes in a facial expression, experimental results demonstrate that the proposed automatic recognition process significantly improves the computational complexity encountered in facial expression recognition while enhancing the ability of recognition