Circuits, Communications and Systems, Pacific-Asia Conference on (2009)
May 16, 2009 to May 17, 2009
With the development of speech recognition technology, there are many speech recognition systems in the world. A non-specific person, small vocabulary, isolated word speech recognition system has been studied in this article. In software the endpoint detection and the adaptive speech recognition arithmetic were used. And at the same time, the TMS320VC5402 is the core of the hardware design. The experimental results show that this system has many advantages such as high recognition precision, high speed and etc. The speech recognition rate rises up to 96%.
Hidden Markov model, Endpoint detection, adaptive, DSP
Zhang Huili, Ren Wenxia, Lv Wenzhe, "Realization of Isolated-words Speech Recognition System", Circuits, Communications and Systems, Pacific-Asia Conference on, vol. 00, no. , pp. 353-355, 2009, doi:10.1109/PACCS.2009.165