摘要
针对基于模板匹配的手指静脉识别过程中存在大量异源匹配导致匹配时间长的问题,提出一种融合哈希粗匹配和加权LTS-HD精确匹配相结合的二次匹配特征识别方法。首先,在特征提取过程自适应设置检测算子的移动步长,减少宽线检测算法中的冗余计算;然后,使用均值哈希算法和差值哈希算法计算静脉图像相似度进行加权融合,利用遗传算法进行权重值寻优,使用融合哈希相似度进行粗匹配识别出部分异源匹配;最后,对于粗匹配无法识别的图像采用加权LTS-HD距离算法进行精确匹配。实验结果表明,自适应移动步长宽线检测法相较于基本宽线检测法的平均特征提取时间降低了44.53%,二次匹配特征识别方法相比于加权LTS-HD方法的平均匹配时间降低了7.93%。
In order to solve the problem of long matching time caused by a large number of heterogeneous matching in the process of finger vein recognition based on template matching,a secondary matching feature recognition method combining hash rough matching and weighted LTS-HD precise matching is proposed in tis paper.Firstly,in the process of feature extraction,according to the distribution of vein points around the current detected pixels,the moving step size of detection operator is adaptively set to reduce the redundant calculation in the wide line detection algorithm.Then,the AHash algorithm and the DHash algorithm are used to calculate the vein image similarity for weighted fusion,and the genetic algorithm is used to optimize the weight value.According to the fusion hash similarity,rough matching is carried out to identify part of the heterogeneous matching.Finally,the weighted LTS-HD distance algorithm is used to accurately match the images that cannot be recognized by rough matching.The experimental results show that the average feature extraction time of the adaptive moving step width line detection method is reduced by 45% compared with the basic wide line detection method,and the average matching time of the secondary matching feature recognition method is reduced by 8% compared with the weighted LTS-HD method.
出处
《工业控制计算机》
2021年第2期73-76,共4页
Industrial Control Computer
基金
国家自然科学基金资助项目(61671285)。
