基于LBP和多层次结构的异质手背静脉身份识别

针对多源异质的手背静脉异质图像的识别研究,提出了基于LBP和多层次结构的识别算法;首先对图像做适当的预处理,然后将LBP特征提取算法编码的手背静脉纹理特征图像作为多层次结构的输入,通过多层次结构的逐层由具体到抽象的特征提取,得到的特征具有更大的鲁棒性;最后该算法在多源异质的手背静脉图像库得到的识别率比传统的算法识别率高,达到96.57%;进一步表明该算法能够较好地解决由于多源异质问题对手背静脉识别所造成的识别率低的影响。

Identity Recognition of Heterogeneous Dorsal Hand Vein Based on LBP and Multi-Layer Structure

Study on identity recognition of the multi-source heterogeneity dorsal hand vein heterogeneity images, recognition algorithm based on LBP and multi-layer structure is put forward. At first, the images are preprocessed, and then the texture feature images of dorsal hand vein coded by the LBP feature extraction algorithm are used as the input of the multi-layer structure, Through layer by layer feature extraction of multi-layer structure from the concrete to the abstract, the features have greater robustness. Finally, the recognition rate of this algorithm is higher than that of the traditional algorithm, and the recognition rate is 96.57%. Further, it shows that the algorithm can effectively solve the problem of the low recognition rate of the dorsal hand vein recognition caused by multi-source heterogeneous problem.