主管：中华人民共和国司法部
主办：司法鉴定科学研究院
ISSN 1671-2072  CN 31-1863/N
中国司法鉴定

中国司法鉴定 ›› 2023 ›› Issue (6): 69-74.DOI: 10.3969/j.issn.1671-2072.2023.06.010

• 鉴定科学 • 上一篇    下一篇

千万人级数据库下相似异源指纹出现情况及排位关系研究

李 硕1，韩文强1，李 康1，2，等   

  1. 1.中国人民公安大学 侦查学院，北京 100038； 2.浙江警察学院 刑事科学技术系，浙江 杭州 310053
  • 收稿日期:2023-02-01 出版日期:2023-11-15 发布日期:2023-11-16
  • 作者简介:李硕（1993—），男，博士研究生，主要从事指纹检验、痕迹检验研究。E-mail:lsppsuc@163.com
  • 基金资助:
    中国人民公安大学拔尖创新人才培养经费支持研究生科研创新一般项目（2022yjsky025）。

Research on the Occurrence of Close Non-Matches in Large-Scale Fingerprint Database and Their Ranking Relationship

LI Shuo1, HAN Wenqiang1, LI Kang1,2, et al   

  1. 1. School of Investigation, People’s Public Security University of China, Beijing 100038, China; 2. Department of Forensic Science, Zhejiang Police College, Hangzhou 310053, China
  • Received:2023-02-01 Published:2023-11-15 Online:2023-11-16

PDF

摘要： 目的 选取指纹区域中最具代表性的斗型纹中心区域，旨在探究千万人级数据库中相似异源指纹的出现情况及其与同源指纹的排位关系，阐述相似异源指纹带来的检验鉴定风险。方法 选取60枚具有代表性的斗型纹捺印样本，对其中心区域进行不同特征组合的标注，发送至指纹自动识别系统进行指纹查询，在前100位候选队列中查找同源指纹及检视相似异源指纹，并统计其排位关系。结果 在900次指纹查询的候选队列中查找到503次同源指纹，检视到474枚相似异源指纹，存在22.6 ％的相似异源指纹排位在同源指纹之前或其队列中没有出现同源指纹。结论 相似异源指纹在当前千万人级指纹数据库中有较高的出现率，且存在一定数量的相似异源指纹排位在同源指纹之前，这将给指纹检验鉴定工作带来较大的风险，需要指纹检验人员在实际工作中提高对相似异源指纹的认识，规避错误鉴定的发生。

关键词: 刑事技术, 指纹自动识别系统, 相似异源指纹, 指纹鉴定

Abstract: Objective This study took the most representative central region of whorl as an example, to explore the occurrence and ranking relationship of close non-matches （CNMs） in a ten million people database, and discussed the inspection and identification risks caused by CNMs. Methods In the experiment, the central region of 60 representative whorl samples was labelled with different feature combinations and submitted to the automatic fingerprint identification system （AFIS） for fingerprint query. The researchers inspected CNMs, searched the same source in the top 100 candidate lists and recorded the corresponding rankings. Results 503 same source and 474 CNMs were found in a total of 900 fingerprint queries. 22.6 % of CNMs were ranked before the same source or the same source did not even appear on the lists. Conclusion CNMs have a high occurrence rate in the current large-scale fingerprint database, and there are a certain number of CNMs ranked before their same sources, which will bring greater risks to fingerprint inspection and identification. Fingerprint examiners are required to improve their understanding of CNMs in actual work to avoid mis-identification.

Key words: criminalistics, automatic fingerprint identification system （AFIS）, close non-matches, fingerprint identification

中图分类号: