Head Pose Estimation of Face: Angle of Roll, Yaw, and Pitch of the Face Image

Abstract

Face recognition has achieved immense success in near-frontal images, but in case of pose variation, it remains an unsolved problem for better efficiency. Many researchers have proposed different approaches for pose-invariant face recognition (PIFR) but still, need to explore more. Pose estimation is useful in pose-invariant face recognition and computer vision. PIFR needs a proper pose estimation to achieve better performance. The head pose estimation can be beneficial for synthesized and reconstruction of the frontal face image. In this paper, we have proposed a geometric approach to estimate the angle of roll, yaw, and pitch of the face image. This estimation will be useful to improve the correlation with frontal images and also it will help to reconstruct a frontal face image. Also, an estimated angle can be used to improve the performance of face recognition..