Designing a system for the recognition of words correct pronunciation by using fuzzy algorithms and multi-layer average method

Abstract

The ability to read is developed in children prior to school. In the other hand ,due to their interest in computer, children spend lots of time playing computer games most of which are useless for them or even harmful .Taking these points in to account, it is essential to design and make the related software. By studying speech recognition algorithms, this study presents an algorithm for recognition of correct pronunciation of Persian words. First, children are asked to look at the picture of a word. Then, they are required to pronounce the word .In the next phase, the word itself is shown and the child is to pronounce the word. For example, first a picture of "bread" is shown and the child is required to name it. Then the word "bread" is shown and child is asked to pronounce it. This process fulfills two objectives: correct pronunciation of Persian words and improving children speed reading. Since children pronounce the words with different accents, differentiation of patterns, pattern recognition and recognition of correct or incorrect speech production need to be clarified.