摘要
通过分析分割算法,结合区域跟踪算法和腐蚀膨胀算法,对基于方向场置信度的分割算法进行了改进;然后,结合一阶对称复数滤波,验证基于传统Poincarê指数法所提取得到奇异点的准确性;在此基础上提出了一种基于“主中心点”脊线跟踪的指纹分类方法,该方法根据“主中心点”附近的脊线信息以及奇异点的数目和相关位置来确定指纹纹型。
In this work, a novel fingerprint classification algorithm is developed, which is based on the ridges tracing of ＂stronger core point＂ （SCP）. This algorithm has many improvements compared to those traditional classificaton methods. First, better performance of an improved f＇mgerprint segmentation algorithm is achieved by using the certainty level of directional fields and region tracing and morphological image processing with respect to the related method proposed in Ref. [9]. Then, a combined singular points detection algorithm is presented which uses one order symmetry complex filtering to proof the validity of the singular points detected by the poincare index. Finally, the type of f＇mgerprint is distinguished according to the ridges＇ information around the SCP, the number and location of singular points.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第8期1752-1756,共5页
Computer Engineering and Design
关键词
方向场置信度
主中心点
脊线跟踪
指纹分类
certainty level of directional fields
stronger core point
ridges tracing
fingerprint classification
