Document Analysis For Machine Translation Of Text Input System Design And Implementation,

With the help of high technology in computers, much convenience for interaction between human and computer has been gained. Document analysis(DA) is concerned with the automatic interpretation of images of printed and handwritten documents, including text, engineering drawings, maps, music scores, etc. Document analysis plays a distinct role in information processing. Research in this area supports a rapidly growing industry and becomes more and more active recentlyThis thesis designs the early part of an automatic machine translation system for document, that is, a document analysis and processing system. It pays attention to the distinct characteristics of document analysis systems when they are used for acquiring text for translation system. To process the input data timely, this thesis based on the traditional document analysis algorithm RLSA and gave out a transforming cut algorithm named SPFIS. It deals with document image data on different resolution and locates connected areas quickly. This helps the system gain a high processing speed. At the same time, the system uses Isodata, a self-organizing data analysis technique, to analyze structures of documents. It solves the problem brought by document type. To adjust to new requirements, traditional document analysis structure is simplified and modified. With its flexible and extendible interfaces, the system can cooperate with newly developed character recognition systems. This realizes the goal of acquiring text information in document images.