指纹图像质量预评价方法研究  

Preprocessing measures of fingerprint image quality

作　　者：赵金辉[1] 硕良勋[1] 曲文斌[1] 

机构地区：[1]石家庄经济学院网络安全实验室,河北石家庄050031

出　　处：《计算机工程与设计》2006年第13期2406-2408,共3页Computer Engineering and Design

摘　　要：指纹图像的质量直接影响着指纹识别系统的可靠性和正确性,如何有效的评估指纹图像的质量和排除低质量的指纹图像是一个比较难的问题。在灰度图像处理的研究基础上,给出了一套基于图像灰度和指纹纹理的指纹图像质量预评价方法。该方法充分考虑指纹识别系统的特点,设立关键参数,并把分步排除与综合评价相结合。该方法速度快、占用内存小,能很好的评估指纹图像的质量,提高系统的可靠性。Fingerprint recognition technology is widely applied as a type of biologic recognition technology nowdays. The quality of the fingerprint image has a direct influence on the reliability and correctness of the fingerprint identification system. The question is how to evaluate the quality of the fingerprint image effectively and how to obviate the low quality of the fingerprint image. Based on the analysis and study of grey image, a method is put forward, which is based on the gray level and texture of the image, to pre-evaluate the image quality. Taking into account characteristic of fingerprint recognition system enough, setting up key parameter, combining step exclusion and synthetic evaluation, this method has the advantage of smaller memory and fast calculating speed, thus the reliability of fingerprint recognition system effectively is improved.

关 键 词：指纹识别 图像质量 方向图 图像纹理 奇异点 

分 类 号：TN911.73[电子电信—通信与信息系统]