指纹匹配算法中采集设备无关性的研究

Research on sensor interoperability in fingerprint matching algorithm

采集设备无关性即指纹识别系统对不同指纹采集器的适应性。基于指纹识别的网络身份认证将面临各式采集器的终端,为推广基于指纹识别的网络认证方法,指纹识别系统必须解决采集设备无关性。文章对指纹特征模式进行研究,提出了适用多种采集器的匹配算法。该算法采用一种基于细节点集的Delaunay三角化的局部结构来调整2个指纹模式,并根据若干匹配成功的局部结构提取出2个模式的比例,进一步将源自不同指纹采集器的指纹模式统一到同一比例。在3款采集器自建的指纹图像库中,指纹交叉匹配结果验证了算法的有效性。

Sensor interoperability means the adaptability of automatic fingerprint identification system （AFIS） to different sensors .The network authentication based on fingerprint identification （NAFI） faces different sensors as terminals .Therefore ,AFIS must aim at sensor interoperability to promote the NAFI .In this paper ,the fingerprint feature patterns are studied and a matching algorithm suit-able for multiple sensors is presented .In this algorithm ,the local structure based on Delaunay trian-gulation（DT） is used to align two fingerprint patterns .Scale of patterns is gained from matched local structures ,and then two fingerprint patterns acquired are unified in the same scale .Three image data-bases using three sensors are established ,and the experimental results show that the presented algo-rithm is competent for cross matching of these databases .