期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2014
卷号：67
期号：3
出版社：Journal of Theoretical and Applied
摘要：The research of face recognition and face distance estimation is a study that is being developed nowadays. This research proposes a model of face recognition by combining wavelet decomposition, PCA and Mahalanobis distance. This method will provide a better of recognition rate and will improve the computation time of face recognition. We also develop the measurement of distance between the face and the camera on the proposed face recognition system. We propose real-time distance estimation on face recognition system using stereo vision camera. We calculate the distance using stereo vision camera based on stereo triangulation method using two lenses. The two lenses of stereo vision camera are used to capture a face image and to determine the 3 dimensional points projected from 2 dimensional points. The calculation of distance is determined by the projection from stereo vision camera. The proposed system is capable to calculate the distance beetween face and camera on the proposed of face recognition. This research also suggests a correlation between recognition rate and the distance estimation beetween face and camera. The recognition rate of the proposed of face recognition can work well in the range of 50-350 cm in real-time conditions.
关键词：Face recognition; face distance estimation; stereo vision; wavelet; PCA; Mahalanobis
