一种融和特征矢量与边缘点特征的掌形认证方法

An Algorithm of Hand Shape Authentication Mixed of Feature Vector and Outline Points Feature

针对常用的两种掌形认证方法(特征矢量法、边缘点特征法)中存在的误受率高、运算量大等问题,提出了手指左、右半轴宽度概念,可将上述两种方法的优点有机结合起来,有效克服了两种方法中各自的缺陷,在与特征矢量法运算量相近的前提下,可以保证误受率和误拒率都较低.由于本方法是针对多个手指进行认证的,可以降低对手指摆放位置的要求,可以取消目前掌形仪中的固定栓,使认证过程更方便和人性化,使未来掌形仪的设计更加灵活.

Aiming at the problems of high FAR and mass calculations in the two common hand shape au thentication algorithm, feature vector method and outline points feature method, this paper puts forward the concepts of finger left half and right half width to synthesize advantages of the two algorithm mentioned above organically and get rid of the disadvantages of them. On the condition of the same calculation as that of feature vector method, FAR and FRR can be decreased. Because the authentication of this algorithm aims at multi fingers, the demand on position of fingers is decreased and fixing poles used in hand shape instrument can be cancelled. So the authentication progress can be more convenient and humane, and the design of future hand shape instrument can be more flexible.