关键词: 生物特征识别 眉毛识别 主成分分析 傅里叶变换 特征提取

Abstract:

This paper analyzes the possibility that human eyebrow works as an independent biometric for personal identification,and proposes the idea that PCA ma  y be applied to eyebrow recognition. Using a small-scale database of eyebrow images taken from 32 persons, the problem of how PCA-based eyebrow recognit ion accuracy is related to the size-normalized methods and the information threshold has been studied respectively in the two cases： one constructing ffeature vectors from pure eyebrow images, the other from their Fourier Transform. The experimental results show that the highest accuracies are respecti vely 60. 00% in the first case and 90. 63% in the second one.

Key words: biometrics, eyebrow recognition, principal component analysis, Fourier transform, feature extrac tion