Speech recognition system
AssigneeNational Cheng Kung University
InventorsJhing-Fa Wang, Jia-Ching Wang, Tai-Lung Chen, Chin-Chan Chang
Life of this patent
Estimating the term…
Abstract
The present invention discloses a complete speech recognition system having a training button and a recognition button, and the whole system uses the application specific integrated circuit (ASIC) architecture for the design, and also uses the modular design to divide the speech processing into 4 modules: system control module, autocorrelation and linear predictive coefficient module, cepstrum module, and DTW recognition module. Each module forms an intellectual product (IP) component by itself. Each IP component can work with various products and application requirements for the design reuse to greatly shorten the time to market.
Claims10
A speech recognition system, designed by the application specific integrated circuit, inputting a speech signal and outputting a recognition result, and using modular design to divide the speech recognition system into 4 modules, and each of the four modules forming an intellectual product (IP) on its won, wherein the four modules including:
- a control module, receiving an external signal for controlling an internal circuit, and said external signal comprises a training button and a recognition button;
- an autocorrelation and linear predictive coefficient module, comprising an autocorrelation unit and a linear predictive coefficient unit, and the two units individually finding the autocorrelation coefficient and the linear predictive coefficient to convert the inputted speech data into the linear predictive coefficient;
- a cepstrum module, using said linear predictive coefficient to output the cepstrum for the speech data model;
- a DTW recognition module, using the dynamic timing warping (DTW) algorithm to compare said speech data model and then outputting a recognition score.
Dependencies are inferred from the claim text.