An introduction to speech recognition

The term 'speech recognition' refers to the technology that enables computers to 'understand' human speech. Recent developments in the speech recognition technology have moved it into the category of credible and useable state-of-the-art technologies. Today, a number of speech recognition systems are available on the market. Various features characterize the potentials and limitations of these systems. A system may be able to understand a small, large, or unlimited set of words. It could be able to recognize the voice of any human speaker, or only a limited number of speakers, and may be fine tuned to deal with over-the-telephone speech. It may even be able to understand various languages. The number of different applications that can be targeted is large. Examples and demo's will show this. There are various ways to develop speech recognition algorithms. Some of the generally used techniques include template matching and statistical models. What speech recognition will do for us in the near future depends a lot on the rate of progress made in the algorithm development and the acceptance of the technology by the general public.