An Eye Feature Extraction Based on Harris Algorithm
In order to improve the recognition rate of the facial recognition system and the efficiency of the algorithm, a method based on Harris algorithm and changes of threshold to extract eye contour automatically is proposed.This method uses the idea of minimum extremal region.It makes use of the dynamic change of threshold form low to high.After location the eye area initially, the location of the eye pupil is located accurately by means of the method of gray statistic.Then Harris algorithm is used to confirm the eye point position.Finally, the characteristics of the eye contour are fitted.The experimental result shows that this method can extract the feature of eye contour effectively and has the better robustness about illumination and facial expression.
feature minimum extremal region harris algorithm expression recognition
Zhao Wen-dong Qi Hui Zhou Hai-yan
Faculty of Computer Engineering Huaiyin Institute of Technology Huaian,China Department of computer Taiyuan Normal University Taiyuan,China
国际会议
成都
英文
873-876
2011-07-15（万方平台首次上网日期，不代表论文的发表时间）