TY - GEN
T1 - Eye detection using intensity and appearance information
AU - Hassaballah, M.
AU - Ido, Shun
PY - 2009
Y1 - 2009
N2 - Eyes are the most salient and stable features in the human face and hence automatic extraction or detection of eyes is often considered as the most important step in face identification and recognition. This paper presents a new method for eye detection of still gray scale images. The method is based on two facts; eye regions exhibit unpredictable local intensity, therefore entropy in eye regions is high and the iris of eye is circle and too dark-compared to the neighboring regions. A score based on the entropy of eye and darkness of iris is used to detect eye center coordinates. Experimental results on two data-bases, namely FERET with variations in views and BiolD with variations in gaze directions and uncontrolled conditions show that the proposed method is robust against gaze direction, variations in views and variety of illumination. It can achieve a correct eye detection rate of 97.8% and 94.3% on the FERET and BiolD images respectively. Moreover, in the case of glasses the performance is still acceptable.
AB - Eyes are the most salient and stable features in the human face and hence automatic extraction or detection of eyes is often considered as the most important step in face identification and recognition. This paper presents a new method for eye detection of still gray scale images. The method is based on two facts; eye regions exhibit unpredictable local intensity, therefore entropy in eye regions is high and the iris of eye is circle and too dark-compared to the neighboring regions. A score based on the entropy of eye and darkness of iris is used to detect eye center coordinates. Experimental results on two data-bases, namely FERET with variations in views and BiolD with variations in gaze directions and uncontrolled conditions show that the proposed method is robust against gaze direction, variations in views and variety of illumination. It can achieve a correct eye detection rate of 97.8% and 94.3% on the FERET and BiolD images respectively. Moreover, in the case of glasses the performance is still acceptable.
UR - http://www.scopus.com/inward/record.url?scp=78449257017&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:78449257017
SN - 9784901122092
T3 - Proceedings of the 11th IAPR Conference on Machine Vision Applications, MVA 2009
SP - 346
EP - 349
BT - Proceedings of the 11th IAPR Conference on Machine Vision Applications, MVA 2009
T2 - 11th IAPR Conference on Machine Vision Applications, MVA 2009
Y2 - 20 May 2009 through 22 May 2009
ER -