A biological speech recognition system by using associative neural networks

Speech recognition has been one of the most challenging subjects in the field of signal processing for years. The recognition of the speech becomes more difficult when there is a noise in the place. Most of the time human beings are able to ignore the environment noise. The secret of this phenomenon is in the ability of human auditory cortex. Therefore knowing about the functionality of human auditory cortex and its implementation can be helpful to improve the quality of a speech recognition system. The presented system in this paper has two parts. The first part filters the input signal and packs it. Then it gets the average of three packets as an identification of the signal and sends it to the second part. This part which is based on the human auditory cortex is an associative neural network that maps the input set to a desired output set. This system is able to recognize a word even a noisy one.