【摘要】 要提高指纹识别的识别率,关键是指纹特征的提取。主曲线是主成份分析的非线性推广,它是通过数据分布“中间”并满足“自相合”的光滑曲线,较好地反映了数据分布的结构特征。本文尝试使用主曲线这种新的方法来提取指纹的细节特征。实验结果表明利用主曲线来提取指纹的结构特征是可行的,它为指纹特征提取的研究提供了一条新途径。更多还原

【Abstract】 The extraction of fingerprint feature is critical to improve the rate of fingerprint recognition. Principal curves are nonlinear generalizations of principal components analysis. They are smooth self-consistent curves that pass through the "middle" of the distribution and they preferably reflect the structural features of the data. In this pa- per, principal curves are used to extract the minutiae of fingerprint. From the result of experiment, it is proved that using principal curves for minutiae extraction is feasible.更多还原