Reader instrument of basic texts to the teaching of blind people

The development of a reader instrument for translating text to speech is presented. The instrument uses commercial components such as a multimedia computer, a flat bed scanner unit, a control and recognition of characters program. The instrument is used to read printed text by blind people. The character recognition is based on a block of preclassification and artificial neuronal backpropagation network. The characters are read through the scanner text from elemental teaching books, and the program locates the words recognized in a base of recorded words to emit them via voice through multimedia systems. The preclassification divides the lower case by its characteristics in the following four groups: (1) acenorsvuxz. (2) bdfhklnˆit. (3) gjpqy. 4) mw. The prototype developed has a limitation; it only works with text books used in elemental education. The type of character is Arial with size of 12 pixels and a performance of 93%