基于功率谱纹理分析的指纹密钥算法  

Fingerprint Key Generation Algorithm Based on Frequency Spectrum Texture Analysis

作　　者：刘慧慧[1] 胡爱群[1] 

机构地区：[1]东南大学信息安全研究中心,江苏南京210096

出　　处：《信息安全与通信保密》2009年第12期70-71,74,共3页Information Security and Communications Privacy

摘　　要：由于生物特征的模糊性,传统的自动指纹识别使用存储特征模板和模糊匹配的方法,存储的模板直接含有大量指纹信息,由此带来模板安全隐患问题。文中探讨在无模板存储和匹配时提取稳定的参数描述,创新性地提出了一种从频谱中提取指纹特征并描述为稳定密钥的方案,借鉴Fourier频谱分析方法从指纹中提取了反映脊线频率信息的环特征和描述方向信息的楔特征。最后提出多次采集整合的改进方案,得到了稳定率82%的密钥参数。For the fuzziness of biological features, the traditional automatic fingerprint identification adopts feature templates saving and fuzzy matching. The saved templates contain fingerprint information, so security problems occur unavoidably. The paper explores how to extract steady features under the condition of with no saving templates and matching, creatively proposes a scheme for extracting features from the frequency spectrum and describes it as steady feature key. Based on Fourier spectrum analysis, the ring feature which represents ridge frequency and the wedge feature which represents the direction information are extracted from the fingerprint. Finally, an improved scheme for multi-time combination is proposed, and its steady percentage of key parameter is up to 82%.

关 键 词：指纹图像 特征提取 功率谱 生物密钥 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]