Read it aloud to me

Abstract

The universal design applied to assistive technologies can help visually impaired person perform some day-to-day tasks as well as everybody. With this aim, the present work focuses on development of photo-to-speech instruments for the visually impaired person. It allows the user to hear text typed on a sheet of paper or written/posted on a wall. To achieve that aim a set of image capture and processing frameworks such as Optical Character Recognition (OCR) and Text to Speech Synthesis (TTS) were integrated. The first versions of the OCR based speech synthesis systems were developed for our native language, Portuguese. A preliminary desktop version was designed under Windows OS, and a version for mobile devices was developed as an application for Android devices. In this paper, we summarize efforts to develop and test a desktop and a mobile version of autonomous photo-to-speech instruments for the visually impaired. The project consisted of integration of selected components, and the CPU applications governing several functionalities: capture of images by the CCD camera; image preprocessing; OCR framework for text recognition; and finally the process of TTS, producing a synthesized voice.