SPEECH RECOGNITION APPROACH TO COMPUTER PROGRAMMING

Abstract

Though, a lot of research has been done in the field of Computer Speech Recognition, it is all related to task specific systems. The language that is spoken to such system is task dependent and can perform only a few task oriented operations. These languages are only for specific tasks and do not meet the requirements of different tasks. At the same time researchers have used the approach of Natural Language Processing to make the machines capable of understanding the natural languages. These systems are very difficult to design because they have to handle a very large vocabulary of the natural language. This makes the search space for the input speech very large, which in turn makes the system very inefficient and inaccurate. At the same time the grammars of natural languages, being not well defined, further complicate the system design. The present work adopts a new approach in which the problems of handling the large vocabulary, and ambiguity of grammar are solved. And, at the same time any operation independent of task can be performed. Firstly, this approach requires a very -small vocabulary to search the proper sequence of words which form the commands of the speech recognition system. Secondly, a well defined grammar is used. This involves computer programming through speech recognition. The recognized commands will be interpreted as a computer program which can perform the desired operation.