• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >A framework for a fast fingerprint identification using a hybrid system.
【24h】

A framework for a fast fingerprint identification using a hybrid system.

机译：使用混合系统进行快速指纹识别的框架。

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

Biometrics is the automated technique of measuring a physical or behavioral characteristic for purpose of recognizing individuality. Physical and behavioral characteristics are defined as things we are and things we do, respectively. Meanwhile, examples of physical characteristics are faces, retina, irises, fingerprints, and hand geometry and examples of behavioral characteristics are handwritten signatures, and voiceprints. Among various Biometrics, the fingerprint identification, which has been widely used for centuries, is much more reliable. Automatic Fingerprint Identification Systems (AFIS) become very popular and are highly desired for new applications such as access control, security application with a relatively small database, as well as criminal identification with a large database. The objective of this thesis is to design a faster and more reliable framework for fully automatic fingerprint identification system. We have focused on several issues in this dissertation. First, a method for estimating local ridge orientation is explored and used in fingerprint enhancement. This information is used for selecting an appropriate enhancement filter. Second, we have proposed an enhancement algorithm that improves the matching accuracy of the AFIS. The proposed enhancement procedure consists of several sequential steps: segmenting of the fingerprint images, determining of local ridges orientation, designing directional linear filters, enhancing the images, and post-processing. The results show good ridge continuity and ridge separation, improve the reliability of minutiae extraction and improve matching performance. Third, a new hybrid fingerprint identification system is presented. Our proposed matching system uses both geometric and statistic approaches. The system consists of fingerprint classification, fast texture matching, and minutiae identification. The experimental results show that our proposed AFIS performs well and the proposed hybrid matching needs fewer steps to identify individuality than traditional methods. The system is practical for a real-time fingerprint application in large database.
机译：生物识别技术是一种自动技术，用于测量生理或行为特征以识别个性。身体和行为特征分别定义为我们是事物和我们所做的事情。同时，物理特征的示例是面部，视网膜，虹膜，指纹和手的几何形状，行为特征的示例是手写签名和声纹。在各种生物识别技术中，已被广泛使用了多个世纪的指纹识别更加可靠。自动指纹识别系统（AFIS）变得非常流行，并且对于诸如访问控制，具有相对较小的数据库的安全性应用程序以及具有大型数据库的犯罪识别之类的新应用程序，都迫切需要它们。本文的目的是为全自动指纹识别系统设计一个更快，更可靠的框架。本文重点研究了几个问题。首先，探索一种估计局部脊取向的方法，并将其用于指纹增强。此信息用于选择适当的增强过滤器。其次，我们提出了一种增强算法，可以提高AFIS的匹配精度。拟议的增强程序包括几个连续步骤：分割指纹图像，确定局部脊的方向，设计定向线性滤波器，增强图像以及后处理。结果表明，良好的山脊连续性和山脊分离效果，提高了细节提取的可靠性，并提高了匹配性能。第三，提出了一种新的混合指纹识别系统。我们提出的匹配系统同时使用了几何方法和统计方法。该系统包括指纹分类，快速纹理匹配和细节识别。实验结果表明，与传统方法相比，我们提出的AFIS表现良好，提出的混合匹配需要更少的步骤来识别个性。该系统适用于大型数据库中的实时指纹应用。

著录项

  • 作者

    Huvanandana, Sanpachai.;

  • 作者单位

    University of Washington.;

  • 授予单位 University of Washington.;
  • 学科 Engineering Electronics and Electrical.
  • 学位 Ph.D.
  • 年度 2002
  • 页码 116 p.
  • 总页数 116
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类 无线电电子学、电信技术;
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号