Implementation of speech recognition on FPGA
- Bangkok : Asian Institute of Technology, 2001
- 84 p.
- Thesis ; no. ME-01-09 .
- Asian Institute of Technology. Thesis ; no. ME-01-09 .
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering, School of Advanced Technologies
Thesis (M.Eng.) - Asian Institute of Technology, 2001
One concept of speech recognition is that the human utterance always contains statistical information, in which can be modeled by applying powerful mathematics. By using Hidden Markov Model approach, the speech recognition system has been implemented successfully. However, the incoming speech signal must be analyzed in the first step before going to any processing. The spectral analysis and the vector quantization methodology are selected for this purpose. Although in the first step, the speech recognition system is implemented properly in software programming but the goal of this thesis is to design an embedded chip that is more useful in the current market. Finally, by using VHDL and EDA tools, the speech recognition is implemented completely on FPGA with recognition efficiency of 92%. The system is speaker-dependent and has a vocabulary containing 52 Thai words as case study.