分布式算术码在指纹识别中的应用研究
发布时间：2018-08-16 16:57
【摘要】：随着科技的快速发展,指纹识别在智能设备上的使用变得越来越普及。因此,人们对指纹数据的安全存储和匹配效率提出了更高的要求。鉴于分布式算术码在数据处理过程中表现出高编码压缩性和高解码准确性的优势,本文将分布式算术码应用于指纹识别,并基于哈希构造存储模型和分布式算术码编码模型创建了指纹信息数据库;基于分布式算术码解码模型提出了反向验证匹配方案,解决了在匹配过程中判定阈值落在模糊区间的匹配问题。主要研究内容与成果如下:(1)利用指纹图案中特征点与特征点相对位置不变且之间指纹纹数不变的性质,建立具有固定格式的指纹数据。在这种表示形式下的指纹数据适合批量处理;便于采用固定的处理形式建立数据库;并且便于通过大量的实验数据结果来验证匹配效率的性能,为本文指纹识别应用提供了数据来源。(2)采用分布式算术码编码数据处理形式和哈希构储结构建立指纹数据模板库。在建立指纹模板数据库时,存在插入重复和查找耗时的问题。本文采用以哈希表作为存储模板,对指纹特征信息采用分步处理(位置数据与信息数据)的方法,并结合哈希变换和分布式算术码编码模型完成数据库的建立;其中,对于在形成位置数据时遇到的插入数据冲撞溢出的问题,研究提出哈希关联函式的方案,并以此解决哈希冲突,从而实现模板数据库数据存储不重复、不遗漏的目的。(3)一般的指纹匹配比对存在一个不能定性的判定区间,当判定结果落在该区间时,比对结果会出现偏差。鉴于此,本研究基于分布式算术码解码模型提出一种指纹数据反向验证方案,提高了数据匹配比对准确性;在解码匹配过程中,可采用算术码码谱辅助解码,提升解码的速度和准确性,进而保证匹配的可靠性和高效性。
[Abstract]:With the rapid development of science and technology, the use of fingerprint recognition in intelligent devices has become more and more popular. Therefore, people put forward higher requirements for fingerprint data security storage and matching efficiency. In view of the advantages of high coding compression and high decoding accuracy of distributed arithmetic codes in the process of data processing, this paper applies distributed arithmetic codes to fingerprint identification. The fingerprint information database is created based on the hash model and the distributed arithmetic code coding model, and the reverse verification matching scheme is proposed based on the distributed arithmetic code decoding model. The matching problem in which the threshold falls in the fuzzy interval in the matching process is solved. The main research contents and results are as follows: (1) the fingerprint data with fixed format is established by using the property that the relative position of feature point and feature point in fingerprint pattern is invariant and the number of fingerprint lines between them is invariable. The fingerprint data in this form is suitable for batch processing; it is convenient to set up a database in a fixed processing form; and it is convenient to verify the performance of matching efficiency through a large number of experimental data. It provides a data source for fingerprint identification applications in this paper. (2) A fingerprint data template library is established by using distributed arithmetic code encoding data processing form and hash structure. In the establishment of fingerprint template database, there are problems of inserting repetition and searching time. In this paper, the hash table is used as the storage template, the fingerprint feature information is processed step by step (position data and information data), and the hash transform and distributed arithmetic code coding model are combined to complete the establishment of the database. For the problem of collision and overflow of insert data in the formation of location data, a scheme of hash correlation function is proposed to solve the hash conflict, so that the data storage of template database can not be duplicated. (3) there is an unqualitative decision interval in the general fingerprint matching. When the result falls in the interval, the result of the comparison will deviate. In view of this, based on the distributed arithmetic code decoding model, this paper proposes a reverse verification scheme for fingerprint data, which improves the accuracy of data matching and matching. Improve the speed and accuracy of decoding, and then ensure the reliability and efficiency of matching.
【学位授予单位】：西北农林科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41
本文编号：2186605
[Abstract]:With the rapid development of science and technology, the use of fingerprint recognition in intelligent devices has become more and more popular. Therefore, people put forward higher requirements for fingerprint data security storage and matching efficiency. In view of the advantages of high coding compression and high decoding accuracy of distributed arithmetic codes in the process of data processing, this paper applies distributed arithmetic codes to fingerprint identification. The fingerprint information database is created based on the hash model and the distributed arithmetic code coding model, and the reverse verification matching scheme is proposed based on the distributed arithmetic code decoding model. The matching problem in which the threshold falls in the fuzzy interval in the matching process is solved. The main research contents and results are as follows: (1) the fingerprint data with fixed format is established by using the property that the relative position of feature point and feature point in fingerprint pattern is invariant and the number of fingerprint lines between them is invariable. The fingerprint data in this form is suitable for batch processing; it is convenient to set up a database in a fixed processing form; and it is convenient to verify the performance of matching efficiency through a large number of experimental data. It provides a data source for fingerprint identification applications in this paper. (2) A fingerprint data template library is established by using distributed arithmetic code encoding data processing form and hash structure. In the establishment of fingerprint template database, there are problems of inserting repetition and searching time. In this paper, the hash table is used as the storage template, the fingerprint feature information is processed step by step (position data and information data), and the hash transform and distributed arithmetic code coding model are combined to complete the establishment of the database. For the problem of collision and overflow of insert data in the formation of location data, a scheme of hash correlation function is proposed to solve the hash conflict, so that the data storage of template database can not be duplicated. (3) there is an unqualitative decision interval in the general fingerprint matching. When the result falls in the interval, the result of the comparison will deviate. In view of this, based on the distributed arithmetic code decoding model, this paper proposes a reverse verification scheme for fingerprint data, which improves the accuracy of data matching and matching. Improve the speed and accuracy of decoding, and then ensure the reliability and efficiency of matching.
【学位授予单位】：西北农林科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41
【参考文献】
相关期刊论文 前8条
1 粘永健;万建伟;何密;辛勤;;基于分布式信源编码的高光谱图像无损压缩研究进展[J];宇航学报;2012年07期
2 杨艳;朱志宇;冯友兵;;将IRA码应用于分布式信源编码中[J];科学技术与工程;2011年34期
3 宋娟;吴成柯;张静;刘海英;;基于分类和陪集码的高光谱图像无损压缩[J];电子与信息学报;2011年01期
4 曾伟超;杨胜天;仇佩亮;;基于随机交织的分布式信源编码的实际设计[J];浙江大学学报(工学版);2009年11期
5 段玉波;佟金鹰;刘丽丽;;一种指纹特征点信息的表示方法[J];佳木斯大学学报(自然科学版);2009年01期
6 赵洪博;刘荣科;;格雷码在分布式信源编码中的应用与性能分析[J];遥测遥控;2008年04期
7 胡琳;姚庆栋;刘云海;;传感器网络中的分布式信息编码[J];信息与控制;2006年02期
8 冯俐俐,李昌禧;指纹中心点的定位和特征匹配方法[J];华中科技大学学报(自然科学版);2002年10期
相关博士学位论文 前3条
1 宋娟;基于分布式信源编码的多光谱图像/视频压缩技术研究[D];西安电子科技大学;2012年
2 薛国栋;分布式信源编码理论与应用研究[D];北京邮电大学;2009年
3 回红;基于结构的指纹表达及其匹配算法研究[D];浙江大学;2002年
相关硕士学位论文 前3条
1 陈亮;改进的分布式算术码研究[D];西北农林科技大学;2014年
2 王阳;基于LDPC码的分布式信源编码研究[D];华中科技大学;2011年
3 马小妹;指纹图像的分类和识别[D];大连理工大学;2002年
，本文编号：2186605
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2186605.html
