Person identification system based on a trapezoid pyramid architecture of a gray-level image
To realize fully automated face recognition, there must be thorough processing from detection of the face in a scene to recognition. There have been many reports on face recognition, however, studies on detection available for recognition are very few. One of the difficulties comes from many variations of input condition such as illumination and background. As for access control systems such as security or login, input conditions can be rather fixed. Under this condition, fully automated person identification by the facial image is tried and achieved. The face in a scene is first sought by coarse-to-fine processing based on a trapezoid pyramid architecture of a gray-level image, and the result is applied to the recognition. The simple algorithm is implemented by software in a personal computer, and this realizes a series of processing within one second.
KeywordsFace Recognition Recognition Rate Facial Image Front View Mosaic Pattern
- 1.M.A.Turk and A.P.Pentland,“Face recognition using eigenfaces”,Proc.CVPR'91,pp.586–591(June '91).Google Scholar
- 3.M.Kosugi,“Robust identification of human face using mosaic pattern and BPN”. In: Neural Networks for Signal Processing II, Proc. of the 1992, IEEE workshop, pp. 296–305 (Sept. '92).Google Scholar
- 5.M.S.Lew and N.Huijsmans,“Information theory and face detection”,Proc.ICPR'96,pp.601–605 (Aug. '96).Google Scholar
- 7.P.J.Burt,“Multiresolution techniques for image representation, analysis and 'smart’ transmission”, SPIE Conf.,Visual Communications and Image Processing IV(Nov.'89).Google Scholar
- 8.M.Kosugi,“Human-face search and location in a scene by multi-pyramid architecture for personal identification”, Systems and Computers in Japan, 26,6, pp.27–38 (June '95).Google Scholar