受限表格识别系统的研究

Constrained Form Recognition System

表格的自动输入、存储、管理已经成为文档智能处理领域的一个重要组成部分。一方面,表格结构复杂多样,使得很难找到一种能良好识别任何表格的通用方法;另一方面,很多领域里使用的表格,需要识别的通常是固定的某些表格单元。因此,设计并实现了一个受限表格识别系统,实验结果表明,该系统提高了表格识别的适应性和正确率。

Inputting,storing and managing forms automatically have already become an important component of the field of intelligent document processing.On one hand,form structures are complicated and various,which make it very difficult to find a general method to recognize all kinds of forms well.On the other hand,in many fields,usually some regular form units need to be processed.So,we develop a constrained form recognition system.Experiments demonstrate that our system enhances the adaptability ,accuracy of form recognition.