基于多模板融合的异质手背静脉身份识别*

针对手背静脉识别过程中的多源异质问题，分别在图像预处理和识别算法上进行改进，提出了自适应手背静脉图像的归一化配准方法，并利用位置信息对SIFT(Scale-invariant feature transform，尺度不变特征变换)特征点进行精确筛选，降低了错误匹配概率，最后采用多模板融合的识别策略对异质图像进行配准识别，通过图像降维，进一步提高识别率和识别效率，使得平均识别结果达到90.17%。与其他算法的对比结果表明，该算法能够较好地解决多源异质问题对手背静脉识别所造成的影响。

Heterogeneity recognition of dorsal hand vein based on multi-template fusion

Aiming at the multi-source heterogeneous problems attack in dorsal hand vein recognition, this paper proposed a self-adaptive dorsal hand vein normalized registration method through improving the image preprocessing method and recognition algorithm. According to the position information, SIFT feature points were accurately selected to reduce the probability of wrong registration. At last, a multi-template fusion recognition strategy was adopted to recognize the heterogeneity images, and the recognition rate and efficiency was further improved through image dimensionality reduction with an average recognition result of 90.17%. By comparing with other algorithms, the proposed algorithm is effective to the recognition of multi-source heterogeneity dorsal hand vein.