基于纹理的计算机笔迹识别算法研究
发布时间:2018-11-28 08:57
【摘要】: 从上世纪60年代计算机辅助笔迹鉴定的研究就开始了。当时,苏联的几名研究者在1966年发表了用计算机进行笔迹鉴定的研究报告,当时他们所用的方法是从俄文字符骨架中提取一些代表点作为特征,如用俄文中的字母“K”作为实验样本,当时得到了75%的鉴定正确率。到了70年代末、80年代初,计算机辅助笔迹鉴定的研究和开发演变成了一个热潮。其中属前联邦德国刑事技术部主持研制的计算机联网笔迹鉴别系统影响最大。该项目有大量专家学者参加了研究,虽然投资大、历时长,但是也获得了一系列行之有效的方法。由于语言文字上的渊源,日本的方法比较适合中国的情况,而且日本学者在这方面发表的文献数目也比较多,质量也都比较高,很值得我们借鉴。 目前中国的文检技术水平在在世界上也是处于领先地位的,特别是在笔迹鉴定这一方面。从上世纪90年代开始,中科院自动化文字识别实验室与北京市刑事科技研究所合作就开始了计算机辅助笔迹鉴定的研究,之后后众多国内的专家学者和研究机构都在致力于笔迹鉴定计算机系统方面的开发,并取得了一些成果,但真正投入广泛使用的到现在还很少。当前,传统的由笔迹鉴定专家人工进行笔迹鉴定的工作方法已难以满足目前工作量很大的笔迹鉴定要求。在计算机技术、互联网及数字图像技术广泛应用的今天,笔迹鉴定工作也应借助科技优势实现数字化、信息化、网络化。因此,我们还需对计算机辅助笔迹鉴定的相关方面进行深入的研究。 本文除了引言以外分为五个大的部分:第一部分是基于纹理的计算机辅助笔迹鉴定算法概述和目前研究状况及问题的提出;第二部分主要简析了计算机辅助笔迹鉴定系统的构成;第三部分是笔迹图像的预处理算法分析;第四部分见简析研究了笔迹特征提取的算法;最后是对基于纹理的计算机笔迹算法的分析。
[Abstract]:The study of computer-aided handwriting identification began in the 1960 s. At that time, several researchers in the Soviet Union published a study on handwriting identification by computer in 1966. The method they used at that time was to extract some representative points from the Russian character skeleton as a feature. For example, using the Russian letter "K" as the experimental sample, 75% accuracy was obtained at that time. By the late 1970s and early 1980s, the research and development of computer-aided handwriting identification had become an upsurge. The computer network handwriting identification system developed by the former Federal Ministry of Criminal Technology in Germany has the greatest influence. A large number of experts and scholars participated in the project. Although the investment is large and the duration is long, a series of effective methods have been obtained. Due to the origin of language and writing, the Japanese method is more suitable for the situation in China, and the number of documents published by Japanese scholars in this respect is also relatively large and the quality is relatively high, which is worthy of our reference. At present, the level of document checking technology in China is also in the leading position in the world, especially in handwriting identification. Since the 1990s, the automatic character recognition Laboratory of the Chinese Academy of Sciences, in cooperation with the Beijing Institute of Criminal Science and Technology, has begun the study of computer-aided handwriting identification. After that, many domestic experts, scholars and research institutes are devoting themselves to the development of handwriting identification computer system, and have made some achievements, but few of them have been put into wide use until now. At present, the traditional manual handwriting identification method by handwriting identification experts has been difficult to meet the current workload of handwriting identification requirements. Nowadays, with the wide application of computer technology, Internet and digital image technology, handwriting identification should also be digitized, informationalized and networked with the help of scientific and technological advantages. Therefore, we need to further study the related aspects of computer-aided handwriting identification. In addition to the introduction, this paper is divided into five parts: the first part is an overview of texture-based computer-aided handwriting identification algorithm and the current research situation and problems; The second part mainly analyzes the composition of the computer-aided handwriting identification system; the third part is the analysis of the pre-processing algorithm of handwriting image; the fourth part is a brief analysis of the handwriting feature extraction algorithm. Finally, the computer handwriting algorithm based on texture is analyzed.
【学位授予单位】:西南政法大学
【学位级别】:硕士
【学位授予年份】:2010
【分类号】:D918.92
本文编号:2362403
[Abstract]:The study of computer-aided handwriting identification began in the 1960 s. At that time, several researchers in the Soviet Union published a study on handwriting identification by computer in 1966. The method they used at that time was to extract some representative points from the Russian character skeleton as a feature. For example, using the Russian letter "K" as the experimental sample, 75% accuracy was obtained at that time. By the late 1970s and early 1980s, the research and development of computer-aided handwriting identification had become an upsurge. The computer network handwriting identification system developed by the former Federal Ministry of Criminal Technology in Germany has the greatest influence. A large number of experts and scholars participated in the project. Although the investment is large and the duration is long, a series of effective methods have been obtained. Due to the origin of language and writing, the Japanese method is more suitable for the situation in China, and the number of documents published by Japanese scholars in this respect is also relatively large and the quality is relatively high, which is worthy of our reference. At present, the level of document checking technology in China is also in the leading position in the world, especially in handwriting identification. Since the 1990s, the automatic character recognition Laboratory of the Chinese Academy of Sciences, in cooperation with the Beijing Institute of Criminal Science and Technology, has begun the study of computer-aided handwriting identification. After that, many domestic experts, scholars and research institutes are devoting themselves to the development of handwriting identification computer system, and have made some achievements, but few of them have been put into wide use until now. At present, the traditional manual handwriting identification method by handwriting identification experts has been difficult to meet the current workload of handwriting identification requirements. Nowadays, with the wide application of computer technology, Internet and digital image technology, handwriting identification should also be digitized, informationalized and networked with the help of scientific and technological advantages. Therefore, we need to further study the related aspects of computer-aided handwriting identification. In addition to the introduction, this paper is divided into five parts: the first part is an overview of texture-based computer-aided handwriting identification algorithm and the current research situation and problems; The second part mainly analyzes the composition of the computer-aided handwriting identification system; the third part is the analysis of the pre-processing algorithm of handwriting image; the fourth part is a brief analysis of the handwriting feature extraction algorithm. Finally, the computer handwriting algorithm based on texture is analyzed.
【学位授予单位】:西南政法大学
【学位级别】:硕士
【学位授予年份】:2010
【分类号】:D918.92
【引证文献】
相关硕士学位论文 前1条
1 贾雪飞;基于纹理强化的微量笔迹鉴别技术研究[D];湖南大学;2012年
,本文编号:2362403
本文链接:https://www.wllwen.com/falvlunwen/fanzuizhian/2362403.html