1、 28 4 S / (1 S ) Vol.28 No.42010 M 7 Journal of Foshan University (Natural Science Edition) Jul. 2010 c I | : 1008-0171(2010)04-0030-05 E L C于昕梅,蒋业文( S / 0 “ , o, 9 . s g J. 9 _ ,2005,22(4):75-77.4ABDOLHOSSEIN Fathi, MOHAMMAD T Manzuri. Eye detection and tracking in video streamsC/Internati-onal Sym
2、posium on Communications and Information Technologies. Sapporo, Japan: W , 2004: 1258-1261.5WANG Peng, MATTHEW B Green, QIANG Ji, et al. Automatic eye detection and its validationC/ /CVPR.San Diego: W , 2005: 164-171.6WANG Jun, LI Jun yin. Detecting and tracking eyes through dynamic terrain feature
3、matchingC/ / Proceedingsof the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Miomi: W , 2005: 78-85.3 I : l Research and complement of locating human eyesYU Xin-mei, JIANG Ye-wen(Department of Electronics and Information Engineering, Foshan University, Foshan 5280
4、00, China)Abstract: In view of the consistence between 2D Gabor function and market acceptance of advancedanimalsnvisual cortex, this paper employs an eye location algorithm based on Gabor transformation inaccordance with the algorithmsndeficiencies: based on the normalization processing to the imag
5、e toreduce the impact of light, and according to image through Gabor transformation having the largeramplitude on the eyes, we get the coordinates of the eyes through the projection method. Thisalgorithm just needs a shorter operating time. By combining with the simple and effective gray-scaleprojec
6、tion, we can greatly improve the speed of location. And the experimental result indicates, thisalgorithm can locate the human eye accurately, and eliminate the impact of non-uniform light, noiseand gadget on the eye location and have stronger adaptability to various face posture changes.Key words:Gabor transformation; eye location; grey-scale projection34 S / (1 S ) 28
