基於Augmented XY-cut之文件影像結構分析

Abstract

本研究從多種類文件中都擁有的排版關係，分析其排版結構，再加上可編寫Rule-Base文件，來達到協助影像分析目的。一般研究方法為了分析文件影像，需對文件影像進行結構分析，好的結構分析結果可以簡化後續理解程序，本研究利用Recursive XY-cut來當作基底，修改成更實用的Augmented XY-Cut分析，Augmented XY-Cut修正Recursive XY-cut只能切到文件欄位的缺點，並加入了空白區塊節點，讓Augmented XY-Cut更符合文件結構，簡化了Rule-Base複雜度。This study use Recursive XY-cut as a base, modified into a more practical Augmented XY-Cut Analysis. General methods for analysis of document images, document image to be on the structural analysis, structural analysis can be simplified a good follow-up to understand. Augmented XY-Cut Fixed Recursive XY-cut field that can only be cut to the shortcomings of the documents and add the whitespace node, so Augmented XY-Cut more in line with the documents structure, simplifies the complexity of Rule-Base analysis.