International Journal of Computer Science Issues

More than a traditional journal...

In this paper We propose a new method able to detecting and correcting errors relating to the recognition of invoice type documents. We rely on automated document readers that can read and recognize the various relevant information in a scanned document. The process on which this method is based consists of digitizing a large volume of documents, and makes them pass through automatic readers of the documents, then carry out the correction of the various errors. The final goal is to find an electronic document reflecting the various information included in the background document. The main goal is the generation of organized electronic documents, like a data basis or files XML; for a specific use. Our approach is based on the language theory through developing a kind of parser which is applicable to the more general case of documents and can easily detect a specific class of errors and correct them.