A voice identification fixed point programming disposal
A processing method and technology of speech recognition, applied in speech recognition, speech analysis, instruments, etc., can solve the problems of small memory, poor recognition performance, slow processing unit execution speed, etc., and achieve the effect of reducing usage
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment Construction
[0020] Referring to Fig. 1, the speech recognition fixed-point processing method of the present invention is as follows:
[0021] S11: providing a voice signal;
[0022] S12: Using an integer algorithm to obtain the cepstral coefficient of the speech signal;
[0023] S13: Establish an acoustic model of a long integer type of the speech signal;
[0024] S14: Using the cepstrum coefficients and the acoustic model to perform the Viterbi algorithm to obtain the identification comparison value.
[0025] The above-mentioned cepstral coefficients include a Mel-Frequency Cepstrum Coefficients (MFCC), and the values stored in the acoustic model of long integer type can be expressed as Mean*32768*2048, Variable*32768*2048, (Transition Probability )*32768 and Gconst*32768, when calculating the identification comparison value, Mean and Variable and Mel cepstral coefficient C M Operation, the result is then calculated with Gconst and Transition Probability, and the C obtained by the a...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 