Classification and image recognition methods in desktop publishing systems

Abstract

Because of the use of extending desktop publishing (DTP) systems in professional publishing, it is very important to realize the automatic recognition errors by input texts and their processing in these systems. The main consideration is made for classification and recognition methods by using word processors (WP) and optical character recognizers (OCR). The classification of errors and an analysis of methods of their recognition is given. It is shown that combining the simple classification and recognition methods gives the same effect as using each complicated method separately. For more complicated texts, with many different fragments and formed commands, the principle of combining classification and recognition methods is even more important. Their realization has to be based on the keyboard input level with vision control of texts according to the principle: 'What you see is what you get' (WYSIWYG). >