摘要
基于具有统计不相关性的最优鉴别变换 ,分析了小样本识别问题 ,提出了抽取人脸图像有效鉴别特征方法 ,在OlivettiResearchLaboratory (ORL)人脸图像库上得到了平均识别错误率为 2 .75 %的实验结果 ,这是目前在ORL人脸图像数据库上所得到的最好实验结果 ,并在南京理工大学NUST60 3人脸图像库上得到平均识别错误率为 0 .9%的实验结果 ,这些结果表明所提出的人脸图像有效鉴别特征方法是有效的。
Based on the uncorrelated optimal discriminant transformation, the recognition problem in the case of a small number of samples was analyzed and a new method was proposed to extract efficient discriminant features for face images. An average error rate of 2.75% was obtained by experiments on Olivetti Research Laboratory (ORL) database,which is the best result on this database up to now. Moreover, an average error rate of 0.9% was obtained with the experiments on the database. These experimental results show that the new feature extract method is very effective.
出处
《南京理工大学学报》
EI
CAS
CSCD
2000年第3期193-198,共6页
Journal of Nanjing University of Science and Technology
基金
国家自然科学基金资助项目! (6 96 72 0 13)
关键词
模式识别
特征抽取
图像处理
人脸识别
pattern recognition, feature extraction, image processing
face recognition
