摘要
基于指纹图像局部区域所固有的方向特性和频率特性,提出了方向场和频率场的概念,设计了一种快速而有效的自适应指纹图像增强算法.该算法基于局部与宏观相结合的原则,不但利用了指纹图像的局部特性,而且结合了局部四邻区域的关联特性.对于指纹图像中脊线方向变化较规则的区域,采用方向为最佳且经过参数优化的单一方向Gabor滤波;对于指纹图像中脊线方向存在突变块的区域,则采用组合的多方向Gabor滤波.将该方法与LinHong方法的增强处理结果图例对比表明,它具有计算量小、增强效果好的特点.
Based on the intrinsic property of orientation and interval of fingerprint textures in a local region of fingerprint images, the concept of the orientation field and the frequency field is proposed. Then, a fast and effective fingerprint enhancement algorithm is designed using dedicated Garbor filter based on the information of orientation field and frequency field in a fingerprint image. The algorithm will not only utilize the information in a local specified region of fingerprint images, but also the associated information of adjacent neighborhood regions with four directions. Based on the combination analyses of specified region and its neighborhood regions, a single Gabor filter with the best orientation direction and other optimal parameters will be selected in the new adaptive fingerprint enhancement algorithm for the regular regions in fingerprint image, and a combination multi Gabor process will be implemented for the non-regular regions, such as the poor image quality regions or the regions with the ridge in abruptly changed direction. The experimental results show that compared with Lin Hong′s algorithms, the algorithm requires less computational processing and provides better enhancement result.
出处
《大连理工大学学报》
EI
CAS
CSCD
北大核心
2004年第5期689-694,共6页
Journal of Dalian University of Technology