Real-Time Head Pose Estimation on Mobile Devices
There are a lot of actual application scenarios of head pose estimation, such as live detection, human interaction, gesture-based equipment, VR devices. This paper proposes a method based on linear regression to get the pose angle. The method utilizes supervised gradient descent method to get the facial feature points, and estimates the head pose by least square regression. In regression modeling, we only use seven feature point positions, which decrease the model size and the computing load. A data normalization technique is employed to eliminate the effect of the camera imaging parameters on pose estimation. Moreover, the training data set is derived from a group of 3D face models according to the camera imaging model, which can provide precise pose parameters. Experiments validate the proposed method, and show it can be run in real time. The algorithm can be easily transplanted to the mobile terminals. We have released our source code at https://github.com/RoboPai/sdm.
KeywordsFeature Point Mobile Terminal Face Model Relevance Vector Machine World Coordinate System
This work was partially supported by the National Natural Science Foundation of China under Grant No. 61572078, and Beijing Municipal Natural Science Foundation under Grant No. 4152028.
- 1.Zhu, X., Lei, Z., Yan, J., Yi, D., Li, S.Z.: High-fidelity pose and expression normalization for face recognition in the wild. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 787–796 (2015)Google Scholar
- 2.Kim, J.B., Kim, H.I., Park, R.H.: Head pose estimation using a coplanar face model for human computer interaction. In: 2014 IEEE International Conference on Consumer Electronics (ICCE), pp. 560–561. IEEE (2014)Google Scholar
- 5.de Farias Macedo, M.C., Apolinário, A.L., dos Santos Souza, A.C.: A robust real-time face tracking using head pose estimation for a markerless ar system. In: 2013 XV Symposium on Virtual and Augmented Reality (SVR), pp. 224–227. IEEE (2013)Google Scholar
- 11.Raytchev, B., Yoda, I., Sakaue, K.: Head pose estimation by nonlinear manifold learning. In: Proceedings of the 17th International Conference on Pattern Recognition, ICPR 2004, vol. 4, pp. 462–466. IEEE (2004)Google Scholar
- 12.Hu, N., Huang, W., Ranganath, S.: Head pose estimation by non-linear embedding and mapping. In: IEEE International Conference on Image Processing 2005, vol. 2, pp. II–342. IEEE (2005)Google Scholar
- 14.Ma, Y., Konishi, Y., Kinoshita, K., Lao, S., Kawade, M.: Sparse bayesian regression for head pose estimation. In: 18th International Conference on Pattern Recognition (ICPR 2006), vol. 3, pp. 507–510. IEEE (2006)Google Scholar
- 19.Xiong, X., Torre, F.D.L.: Supervised descent method and its applications to face alignment, vol. 9, pp. 532–539 (2013)Google Scholar