Home » Product » Real Time Speech Recognition in Matlab Using HMM/KNN with Wavelet, LPC and Wavelet+LPC method

Real Time Speech Recognition in Matlab Using HMM/KNN with Wavelet, LPC and Wavelet+LPC method


Real Time Speech Recognition in Matlab Using HMM/KNN with Wavelet, LPC and Wavelet+LPC method
$9.99
  • Description

 

1. Download and Unzip the folder

2. Open Matlab ( 2012 or later version)

Inline image 1

3. Type MainGUI  in the Matlab command prompt

4. When GUI Appears, Click on TrainLPC and Select any File from StandardTrainingDatabase folder

Inline image 2

5. Similarly Click on Train Wavelet and Train Combined

This will complete the training of all the methods

6. Now for Testing Click on Test LPC with HMM and Select a file from StandardDatabaseForTesting

Inline image 3

The objective here is to recognize the sound that you are giving as input. Say if you give 3__2,

project should show ’3′ if not, it’s a mis detection.  Similarly you can test the other options.

============================================================

So far what you did was all for standard training and testing for speech recognition. Now we will extend

the method for real time. For real time you must have microphone else accuracy will be low.

7. For real time Training click “Generate Database Sample” ( before that delete number 0 in text box and reenter)

Inline image 4

You must generate samples for 0,1,..9 after every sample completes it shows a message

Database Updated.

After that you can go inside RealTimeDatabase folder. you will see files like 0__1,0__2…9__1 etc

8. Time for Real time Testing

 Click on Speak and Recognize (LPC) and speak the word you want to recognize quickly.

Note: Many features are lockein Demo Version which you will get with the code.

Facebook