Keywords: cross ratio, face detection, face verification, feature extraction, geometric invariance
Received January 1, 2005; accepted June 30, 2005.
Communicated by C. C. Chiang, H. Y. Liao and K. C. Fan.
| Previous | [ 1] | [ 2] | [ 3] | [ 4] | [ 5] | [ 6] | [ 7] | [ 8] | [ 9] | [ 10] | [ 11] | [ 12] | [ 13] | [ 14] | [ 15] |
¡@
Li-Wei He, Tsorng-Lin Chia and Chen-Kuei Yang
Department of Computer and Communication Engineering
*Department of Computer Science and Information Engineering
Ming Chuan University
Taoyuan, 333 Taiwan
E-mail: {tlchia; ckyang}@mcu.edu.tw
A new, fast method for verifying human faces using geometric invariants is proposed.
The proposed method takes advantage of the properties of geometric invariants to
deal with background changes and solve illumination, translation, scaling, and rotating
problems. The method includes locating and segmenting the facial image, extracting the
facial features, and implementing recognition processes. Only 25 sets of cross ratios are
needed to compare faces in the matching process in order to identify whether the query
image exists in the database or not. The computational cost of the proposed method is
significantly reduced in both processing and matching time for real time application. In
addition, the proposed method achieves good performance and invariance even under
external environmental changes. Experimental results are given to show that the proposed
method is feasible and effective for real time application to face detection.
Received January 1, 2005; accepted June 30, 2005.
Communicated by C. C. Chiang, H. Y. Liao and K. C. Fan.