Pose Invariant Face Recognition Method Using Stereo Vision
The paper presents a method to decrease dependency of image-based (appearance-based) face recognition solutions, such as Eigenfaces method, on a head orientation. Using information of a human head shape, recovered by stereo vision technique, a normalized face image (en face view) is synthesized. The normalized image is then processed by an image-based recognition subsystem. To effectively deal with large head rotations, where large area of the face is invisible due to occlusion, Eigenfaces recognition method is extended to operate on half-face images. Presented technique is verified experimentally to prove that it can be used to increase performance of image-based face recognition methods in applications where head pose of the subject is not controlled.
KeywordsFace Recognition Face Image Stereo Vision Head Orientation Virtual Camera
Unable to display preview. Download preview PDF.
- 2.Fehn, C.: Depth-Image Based Rendering (DIBR), Compression and Transmission for a New Approach on 3D-TV. In: Stereoscopic Displays and Virtual Reality Systems XI, Proceedings of the SPIE, vol. 5291, pp. 93–104 (2004)Google Scholar
- 3.Hirschmuller, H.: Accurate and Efficient Stereo Processing by Semi-Global Matching and Mutual Information. In: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition CVPR 2005, vol. 2, pp. 807–814 (2005)Google Scholar
- 6.Schreer, O., Kauff, P., Sikora, T.: 3D Videocommunication, Algorithms, concepts, and real-time systems in human centered communication (2005)Google Scholar
- 7.Telea, A.: An image inpainting technique based on the fast marching method. Journal of Graphics, GPU, and Game Tools 9, 23–34 (2004)Google Scholar
- 8.Turk, M.A., Pentland, A.P.: Face Recognition Using Eigenfaces. In: Proceedings of Computer Vision and Pattern Recognition, CVPR 1991, pp. 586–591 (1991)Google Scholar
- 10.Zhao, W., Chellappa, R.: Image-based Face Recognition: Issues and Methods. In: Javidi, B. (ed.) Image Recognition and Classification, Marcel Dekker, New York Basel (2002)Google Scholar