摘要
人脸特征是最自然直接的生物特征,它具有直接、友好、方便的特点,易于为用户接受。人脸识别由于其在监控、罪犯识别、人机交互等方面广泛潜在的应用,已成为图像处理、模式识别和计算机视觉等学科最活跃的研究领域。线性鉴别分析是特征抽取中最为经典和广泛使用的方法之一。近年来,在小样本情况下如何抽取Fisher最优鉴别特征一直是许多研究者关心的问题。文中阐述了应用Fisher判别法在人脸图像样本分类方面的运用。在标准数据库ORL人脸库和Yale人脸数据库上仿真的试验结果证实了方法的有效性和稳定性。
Compared to other biological characteristics by using the characteristics of human face is the most natural and direct mean, as it has straightforward, friendly, convenient features, and easy for users to accept. Due to Face recognition＇s potential and wide range of applications in control, criminal identification, and human-computer interaction, it has became the most active area of research in the image processing, pattern recognition, computer vision and other subjects. In recent years, how to extract Fisher optimal features in the situation of small samples has been a concern of many researchers.This paper highlighted the application of Fisher discrimination method in the use of face image samples classification. Experimental results on ORL face database and Yale database show that the proposed method is efficient and robust.
出处
《电子设计工程》
2012年第24期178-180,共3页
Electronic Design Engineering
关键词
人脸识别
Fisher判别法
小样本问题
样本分类
face recognition
Fisher discriminant analysis
small sample size problem
sample classification
