Automatic Speech Recognition

Abstract

For many years, speech recognition by machine existed only in the minds of science-fiction writers. However, in recent years, the real problem of automatic speech recognition (ASR) has been addressed in many research laboratories throughout the world. The ultimate goal of this research is to produce a machine which will recognise accurately normal human speech from any speaker. Such a machine could be used in a wide variety of applications including speech input to computers, office automation, factory automation, security systems, aids for the handicapped, consumer products and almost any job involving hands-busy/eyes-busy activities. In an ever-expanding information technology and advanced telecommunications age, speech recognition and synthesis devices would permit remote access to a wide variety of information services over the telephone.