An Embedded 3D Face Recognition System Using a Dual Prism and a Camera
In this paper, a single camera and a dual prism are integrated to implement a three-dimensional face recognition system. The proposed system is implemented on an embedded development platform named UBIKIT6612. A dual prism placed in front of the camera is used to simulate human binocular vision. We then used the active appearance models (AAM) to find out the corresponding feature points and calculate the depth of the face by stereo vision. Accordingly, three-dimensional facial model of each member is constructed. Facial features extracted from the 3D facial models are used for identification. To promote the recognition accuracy, we first exclude most of non-members by support vector data description (SVDD), followed by conducting a multi-class support vector machines (SVM) for face recognition. Experimental results show that the proposed method of the exclusion of non-members works more efficiently than those of traditional methods.
KeywordsThree Dimensional Face Recognition Active Appearance Model
Unable to display preview. Download preview PDF.
- 1.Vijaya Kumar, B.V.K., Savvides, M., Venkataramani, K., Xie, C.: Spatial frequency domain image processing for biometric recognition. In: Proceeding of IEEE ICIP, vol. 1, pp. 22–25 (September 2002)Google Scholar
- 2.Shimizu, M., Yoshizuka, T., Miyamoto, H.: A gesture recognition system using stereo vision and arm model fitting. International Congress Series, vol. 1301, pp. 89–92 (2007)Google Scholar
- 3.Chang, C.Y., Huang, C.S.: Application of active appearance model for dual-camera face recognition. In: Proceeding of International Conference on Information Security and Intelligence Control, pp. 333–336 (2012)Google Scholar
- 6.Bouguet, J.Y.: Camera Calibration Toolbox for Matlab, http://www.vision.caltech.edu/bouguetj/index.html
- 10.Chang, C.C., Lin, C.J.: LIBSVM: A library for support vector machines (2001), http://www.csie.ntu.edu.tw/~cjlin/libsvm