Conversion of non-audible murmur to normal speech through Wi-Fi transceiver for speech recognition based on GMM model

摘要

Capturing of speech information like non-audible murmur or whispered voice from human body using NAM microphone(Non-audible murmur) and transmits via Wi-Fi transmitters and focus on its application in automatic speech recognition systems. The NAM microphone serially connected to Wi-Fi transceiver is attached behind the murmured human ear, and captures the quietly uttered murmur. It is applicable in automatic speech recognition systems, where privacy is important in human-machine communication. Moreover the NAM sensor and microphone receives the speech signal directly from the body and transmits to the recognition system using Wi-Fi transmitters, provides robustness against the environmental noises. The application of speech processing is speech recognition, speech analysis, speech transform and can be used by sound impaired people. By providing a vocabulary dictation system the murmured voices will be recognized and can provide an accuracy of speech with an analysis.