指纹特征点区域分布规律应用研究  

Application of fingerprint feature point distribution regularity in fingerprint comparison

作　　者：张宇[1] 吴春生[1] 孙忠[1] 冯才刚[1] 

机构地区：[1]北京市公安局刑事侦查总队,北京100054

出　　处：《刑事技术》2014年第6期7-10,共4页Forensic Science and Technology

基　　金：北京市科技计划(No.Z121100000312099)

摘　　要：目的针对指纹自动识别系统评分机制所存在的局限性,探讨根据指纹特征点分布进行加权的方法。方法通过总结大量人群指纹特征点区域分布规律,按照分布密度在计算中进行加权。结果在对反查指纹加权计算的试验中,同一指纹加权后的分数相比其它指纹得到了提高。结论采用此种加权机制,对于现有基于特征点的指纹自动识别系统准确率提高是有效的。Objective To analyze the limitation of scoring mechanism used in automatic fingerprint identification system and explore a weighted method based on fingerprint feature point distribution. Methods A large group of fingerprints （12 million people） were studied and distribution regularity of fingerprint feature points was summarized. According to the density distribution of fingerprint feature points, the weighted factors were added into the score. Results The test matching was conducted and the results demonstrated that the weighted score of same fingerprint compared to other fingerprint was improved. Conelu- sions This weighted scoring mechanism based on fingerprint feature point distribution is effective for accuracy improvement of automatic fingerprint identification system.

关 键 词：指纹 特征点 分布规律 比对 

分 类 号：D918.91[政治法律—法学]