针对手指长度具有可测量的特性,提出了一种基于手指相对长度进行身份识别的方法。首先通过曲率的方法定位特征点,依据特征点计算手指的绝对长度;然后应用绝对长度间的相对长度构成特征向量;最后计算特征向量间的欧式距离,实现身份识别。实验表明,通过人工测量254个人的手指长度,利用手指的相对长度进行身份识别可以达到80.86%的识别率;从中选择180个人进行试验,应用本文的特征提取和匹配算法可以达到78.52%的识别率,识别效果接近实际测量值。
A method for personal identification is presented based on relative finger length according to that finger length is measurable. In this method, feature points are localized by using curvature, according to which the absolute length of fingers can be calculated, then an eigenvector composed of relative lengths can be obtained. These relative lengths are generated from the absolute length of fingers. At last, personal identification is achieved by computing Euclidean distance between eigenvectors. In the experiment, finger lengths of 254 persons are manually measured and the accurate identification rate is 80. 86 by using relative finger length. When the proposed arithmetic is used to identify 180 persons of them,the accurate identification rate can reach 78.52 %. These experimental results show that personal identification can Be achieved by using the relative finger length and the effect of the proposed arithmetic is closed to the manually measurement.