无方向的三角形匹配指纹识别

发布时间：2018-11-22 17:33

【摘要】：目的指纹匹配是自动指纹识别系统研究的核心内容之一,匹配算法的好坏直接影响识别系统的效能。目前,大多数点模式匹配算法都依赖于指纹方向场的求取,由于输入的指纹图像存在平移、旋转和尺度变化,因此同一个手指在不同时间获得的指纹图像的方向场是不同的,这不仅增加了计算量,也影响了指纹识别的精度。针对上述问题,提出了无方向的三角形匹配算法。方法提出的三角形匹配算法是以平面中任意点与一个确定的三角形之间的位置结构稳定性为理论基础的。首先,分别在待识指纹图像和模板指纹图像中确定基准三角形;其次,将各个特征点与基准三角形三个顶点的距离组成有序三数组;最后,利用数组的相等程度对指纹相似度进行匹配判断。结果采用国际标准测试库FVC2004进行综合性能比对实验,实验结果表明,与其他几种匹配算法相比,本文方法在识别精度上提高了27.97%33.81%,在比对时间上降低了3%5%,在不同旋转角度下误匹配率平均降低了约86.63%,对噪声、平移、旋转和形变有足够的适应能力,具有较高的容错能力和鲁棒性。结论无方向的三角形匹配算法是一种全局模式的算法,该算法不受指纹图像方向及其位置的影响,实现过程简单,识别精度高,平均比对时间少,适用于处理不同类型的图像数据。
[Abstract]:Objective fingerprint matching is one of the core contents of automatic fingerprint recognition system (AFIS), and the effectiveness of the system is directly affected by the matching algorithm. At present, most point pattern matching algorithms depend on the direction field of the fingerprint. Because of the translation, rotation and scale change of the input fingerprint image, the direction field of the fingerprint image obtained by the same finger at different times is different. This not only increases the computational complexity, but also affects the accuracy of fingerprint recognition. In order to solve the above problems, an undirected triangle matching algorithm is proposed. The algorithm is based on the stability of the position structure between any point in the plane and a certain triangle. Firstly, the datum triangle is determined in the fingerprint image and template fingerprint image respectively; secondly, the distance between each feature point and the three vertices of the base triangle is formed into an ordered triple-array. Finally, the similarity of fingerprint is judged by the equal degree of array. Results the comprehensive performance ratio experiment was carried out by using international standard test library FVC2004. The experimental results show that compared with other matching algorithms, the recognition accuracy of this method is increased by 27.973.81, and the comparison time is reduced by 35%. The mismatch rate at different rotation angles is reduced by about 86.63 on average. It has sufficient adaptability to noise, translation, rotation and deformation, and has high fault tolerance and robustness. Conclusion the algorithm of triangle matching without direction is a kind of global pattern algorithm. The algorithm is independent of the direction and location of fingerprint image, and has the advantages of simple realization, high recognition accuracy and less average time. Suitable for processing different types of image data.
【作者单位】：内江师范学院四川省高等学校数值仿真重点实验室;内江师范学院四川省数据恢复重点实验室;
【基金】：国家自然科学青年基金项目(11502121) 四川省高校创新团队(13TD0001) 四川省教育厅重点项目(16ZA0310) 内江师范学院校级重大成果转化项目(14CZ02)~~
【分类号】：TP391.41

【相似文献】