Eye localization based on hue image processing

Author

Wang, Wen-Tao ; Xu, Chao ; Shen, Hong-Wei

Author_Institution

Peking Univ., Beijing

fYear

2007

fDate

Nov. 28 2007-Dec. 1 2007

Firstpage

730

Lastpage

733

Abstract

A new eye localization method based on hue image processing is proposed in this paper. First, face extraction method based on skin-color information is improved by forehead skin localization and lip detection. Second, vertical projection in hue image is used instead of horizontal projection in intensity or edge images in other works. At last, the eye center is located by a peak value extraction algorithm. This method is time efficient as no gradient or other time consuming computations are used. The experiment result shows that a 320times240 color image needs 8 ms to locate the centers of both eyes on a PC with an AMD Sempron ^TM 2500+ at 2.25 GHz CPU. The average detection rate is about 95.25%.

Keywords

feature extraction; image colour analysis; eye localization; face extraction method; horizontal projection; hue image processing; peak value extraction algorithm; skin-color information; vertical projection; Color; Convolution; Data mining; Eyes; Face detection; Filters; Image edge detection; Image processing; Infrared detectors; Skin;

fLanguage

English

Publisher

ieee

Conference_Titel

Intelligent Signal Processing and Communication Systems, 2007. ISPACS 2007. International Symposium on

Conference_Location

Xiamen

Print_ISBN

978-1-4244-1447-5

Electronic_ISBN

978-1-4244-1447-5

Type

conf

DOI

10.1109/ISPACS.2007.4445991

Filename

4445991