基于快速遗传算法的指纹匹配  

Fingerprint Matching Based on Fast Genetic Algorithm

作　　者：张圆圆[1] 景晓军[1] 

机构地区：[1]北京邮电大学信息与通信工程学院,北京100843

出　　处：《计算机工程》2011年第24期13-15,共3页Computer Engineering

基　　金：国家自然科学基金资助项目(60872148)

摘　　要：针对标准遗传算法(GA)存在收敛速度慢且容易早熟的现象,提出一种快速GA算法,将其应用于全局点模式指纹匹配中。该算法通过在迭代过程中逐步缩小种群规模与优秀个体数目,降低突变概率,并尽量选择适应值高的个体,使匹配过程快速收敛于全局最优解。实验结果表明,该算法能提高指纹匹配的速度和准确率。A global point pattern fingerprint matching method based on Genetic Algorithm（GA） is presented. In order to avoid the low efficiency and premature convergence of standard GA, an fast GA is proposed. The algorithm can make the matching process converge rapidly to the global optimal result through narrowing down the population size and the number of best individuals step by step, reducing the probability of mutation, and trying to select individuals with high fitness. Experimental results show that the improved algorithm can enhance the accuracy of fingerprint matching with less time cost.

关 键 词：指纹匹配 指纹预配准 点模式比对 遗传算法 模拟退火算法 

分 类 号：TP306[自动化与计算机技术—计算机系统结构]