基于有向图的流式文档逻辑结构重构方法

摘要

To solve the shortcomings of logical structure reconstruction of re-flowable document,the algorithm based on a directed graph was put forward to improve the previous algorithms.The logical structure was extracted from the template document,and the directed graph was constructed by sieving logic tags of document to be searched.In addition,the single source shortest path algorithm was applied to filter out redundant logic tags to get the optimal document logical structure.The results show that the algorithm can effectively improve the accuracy of logical structure recognition,which forms the foundation of document content understanding,automatic typesetting and format checking,etc.%为解决以往流式文档中逻辑结构重构方法的不足，提出一种基于有向图的流式文档逻辑结构重构方法。从模板文档中提取逻辑结构，在此基础上，通过对待查文档中的逻辑标签进行筛选建立有向图，利用有向图的单源最短路径算法过滤多余的逻辑结构，得到最佳的文档逻辑结构。实验结果表明，该算法能够有效改善逻辑结构识别的准确性，对于文档内容理解和自动化排版与格式检查等应用具有重要意义。