西安电子科技大学--生命科学技术学院
 
当前位置: 主页 > 研究成果 > 代表性成果 >

18、基于蚁群优化的形变指纹匹配算法

更新时间:2015-08-01 14:47 点击:
  

 形变指纹匹配是指纹识别中的一个难点问题,针对严重形变指纹中细节点对应关系很难确立的问题,我们提出了一种新的基于蚁群优化的细节点匹配算法。首先通过细节点局部方向信息和局部细节点结构来衡量细节点相似度,通过局部匹配构建分配图,从该图中人工蚂蚁可以有效的构建出可行解;然后通过伪贪婪法则在邻域内寻找细节点对应关系并推广到细节点邻域范围内;最后选择匹配分数最大的解作为最后的结果。在FVC2004 DB1 和自建的FINGERPASS 交叉匹配数据库上的实验结果证明了我们的算法可以有效的提高形变指纹匹配的精度,对来自于不同采集模式的指纹图像效果尤为突出。

A novel ant colony optimization algorithm for large-distorted fingerprint matching 
Large-distorted fingerprint matching is still a challenging problem. We proposed a novel ant colony optimization algorithm to establish minutiae correspondence in large-distorted fingerprints. Firstly, minutiae similarity is measured by local features, and an assignment graph is constructed by a local search from which the artificial ants can effectively construct a feasible solution. Then, the minutiae correspondences are established by the pseudo-greedy rule and local propagation. Pheromone matrix is updated by a local update rule and a global update rule. Finally, the minutiae correspondences that maximize the matching score are selected as the final result. Experiments on FVC2004 DB1 and FINGERPASS cross-matching database that was established by our lab demonstrate that the proposed algorithm effectively improves the performance of large-distorted fingerprint matching, especially for those fingerprint images acquired from different modes of acquisition.
 

刚性匹配与蚁群优化匹配算法在交叉匹配数据库上的对比:(a)由刚性配准算法得到的结果.(b)由蚁群优化算法得到的结果。其中红色圆形表示正确的细节点对应关系,蓝色正方形表示错误的细节点对应关系。

Comparison of minutiae correspondences obtained by rigid registration and ACO from a cross-matching database. (a) Minutiae correspondences obtained by rigid registration, (b) minutiae correspondences obtained by Algorithm ACO. Red circles denote genuine minutiae correspondences while blue squares denote false minutiae correspondences.
 
 
------分隔线----------------------------
西安电子科技大学 生命科学技术学院 版权所有 2009-2016
电子工程学院网络信息中心制作维护  管理员信箱 xidianlife@163.com