Face Recognition Technology for Robot Vision

摘要

Face recognition is an important function of robot vision. Face recognition implemented on a robot requires robustness to changes in facial expression and facial direction, because the robot cannot control these changes.Toshiba has developed a face recognition technology that satisfies these requirements using sequential images. Our method can accurately recognize a face based on information on the three-dimensional shape from the sequential images. Experimental results have verified that our method is robust in comparison with the conventional methods using a single static image.