基于复合梯度向量的指纹匹配算法

发布时间：2019-06-14 14:35

【摘要】：针对接触式指纹识别中存在的非线性形变问题,提出了一种基于复合梯度向量(Composite Gradient Vector,CGV)的指纹匹配算法.该算法首先在经过预处理的指纹图像上建立直角坐标系和第一个基向量,并对指纹图像进行旋转;然后在旋转后的指纹图像上采样基向量,组建向量簇,提取极大梯度向量,并以极大梯度向量为元素组建复合梯度向量;最后以分层标记规则对复合梯度向量进行分层标记,通过对指纹库进行指纹检索、复合梯度向量匹配、维度和梯度匹配,识别出指纹图像.实验结果表明,该方法克服了接触式指纹识别中旋转、偏移、拉伸等问题,有较强的抗非线性形变能力,同时具有较快的识别速度和较高的识别准确率.
[Abstract]:A fingerprint matching algorithm based on composite gradient vector (CGV) is proposed for non-linear deformation in contact fingerprint identification. the method comprises the following steps of: firstly, establishing a rectangular coordinate system and a first base vector on a pre-processed fingerprint image, and rotating the fingerprint image; and then sampling a base vector on the rotated fingerprint image, forming a vector cluster, and extracting a great gradient vector, And a composite gradient vector is formed by taking a maximal gradient vector as an element, and finally, the composite gradient vector is subjected to a hierarchical mark by a hierarchical marking rule, and the fingerprint image is identified by carrying out fingerprint retrieval on the fingerprint library, matching the composite gradient vector, the dimension and the gradient. The experimental results show that the method overcomes the problems of rotation, offset, stretching and the like in the contact fingerprint identification, and has strong anti-non-linear deformation capability, and has faster identification speed and higher recognition accuracy.
【作者单位】：辽宁工程技术大学工商管理学院;辽宁工程技术大学软件学院;
【基金】：国家自然科学基金(No.61172144)
【分类号】：TP391.41

【相似文献】