文档矢量化技术的研究进展与应用  

Research Progress and Application of Document Vectorization Technology

作　　者：王彤 陆利坤 

机构地区：[1]北京印刷学院信息工程学院,北京

出　　处：《图像与信号处理》2024年第4期416-426,共11页Journal of Image and Signal Processing

摘　　要：文档矢量化是一种将文档内容转化为数学向量表示的技术,一般来说就是将光栅图像或者栅格图像转换为矢量图像。通过矢量化,可以将文本数据转化为计算机可以理解和处理的形式,从而将文档资料通过计算机矢量化的格式(例如OFD,PDF等)完整地保存下来,为印刷过程中的文本处理、信息检索等领域提供了更多可能性。首先,介绍了文档矢量化的背景;其次,介绍了传统文档矢量化模型;然后,将传统方法到基于深度学习的方法进行了全面综述并对不同的方法进行了比较;最后,对文档矢量化的应用领域和发展进行探讨和展望。Document vectorization is a technique that converts the content of a document into a mathematical vector representation, generally a raster image or raster image into a vector image. Through vectorization, the text data can be converted into a form that the computer can understand and process, so that the document data can be completely saved through the computer vectorized format (such as OFD, PDF, etc.), providing more possibilities for text processing, information retrieval and other fields in the printing process. Firstly, the background of document vectorization is introduced. Secondly, the traditional document vectorization model is briefly introduced. Then, the vectorization and the key techniques of vectorization processing in recent years are introduced. Finally, the application fields and development of document vectorization are discussed and prospected.

关 键 词：文档矢量化 矢量图像 深度学习 自然语言处理 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]