Nalezení textu u JPEG obrazu a jeho následná audio syntéza v českém jazyce

Abstract

In my bachelor thesis I deal with problematic of getting the text from the JPEG image and subsequent audio synthesis of acquired text in Czech language. Procedure for solving the above problems can be separate into four parts. In the first part is performed selected method of segmentation on the input JPEG image for improve quality of text in image. In the second part is extracted text from the JPEG image by using the method of optical character recognition (OCR). In the third part is the acquired text edited into appropriate shape for audio synthesis, to give us the best possible result. In the fourth part is performed itself audio synthesis of edited text in Czech language. In the end of thesis I evaluate acquired results with regard of quality of resulting audio synthesis. The application is targeted for visually impaired users, which should facilitate the acquisition of textual informations. Selected programming language for this work is Java language.