A Computer-Based System to Support Forensic Examination of Japanese Documents

这个研究项目已经进行了两年多（2001-2002）。这项研究项目的目的是发展一个以计算机为基础的系统，以支持对日本文件进行法医检查。我们能够达到这个研究项目的目的。研究项目所获得的成果总结如下。(1)我们开发了一个手写对比图表的交互式准备系统，以支持日本法医文书分析。本系统由三大处理模块组成；(1)手写体提取模块，用于从文档中分割组件手写体；(2)手写体字符识别模块，用于识别分割后的手写体，以便高效地制作图表；(3)图表制作模块，用于显示疑问文档和已知文档的同类手写体，方便对比。每个模块与用户交互工作，并且是菜单和鼠标驱动。在WindowsXP操作系统下，使用c++ Builder实现了原型系统。通过一些案例研究，证实了该系统在编制图表方面比传统方法更快、更有效。我们开发了一种半自动提取笔迹特征的计算机算法。提取的特征是笔画的结构，如笔画的长度比、两笔画的相交角等。这些是许多特征中最重要的特征，文件审查员用来检查日本的笔迹。为了验证笔迹的个性，我们检查了这些特征在一个人的笔迹样本和不同人的笔迹样本之间的变化。通过分析，我们能够验证笔迹的个性。在系统中加入特征分析功能，完善系统的配套功能是今后的工作。

This research project has been carried out over two years (2001-2002). The aim of this research project is to develop a computer-based system to support forensic examination of Japanese documents. We were able to achieve the aim of this research project. The summary of results obtained by the research project is as follows.(1) We have developed an interactive preparation system of handwriting comparison charts to support Japanese forensic document analysis. This system consists of three major processing modules ; (1) handwriting extraction module for segmenting component handwritings from documents, (2) handwritten character recognition module, which recognizes segmented handwritings for efficient preparation of charts, (3) chart preparation module, which displays the same kinds of handwritings of both questioned and known documents for easy comparison. Every module works interactively with user, and is menu and mouse driven. The prototype system has been implemented using C++Builder under WindowsXP operating system. Though some case studies, it was confirmed that this system is faster and more effective than traditional method in preparing charts.(2) We have developed a computer algorithm, which extracts features of handwriting semi-automatically. Extracted features were stroke structures such as length ratio of strokes, crossing angles between two strokes, etc. These are most important features of many features, which document examiners have used for examining Japanese handwritings. To validate individuality of handwriting, we examine the variation of these features within a person's handwriting samples and between the handwriting samples of different people. Through the analysis, we were able to validate the individuality of handwriting.It is the future work to add the feature analysis function to this system to improve the supporting function of this system.