Speech Recognition Engines

Flat
IMPORT ACOUSTIC MODEL TO HTK
User: yacine87
Date: 8/21/2017 10:07 am
Views: 3048
Rating: 0

i have two (02) that represente an acoustic model in htk  "models" ; "macros"  above samples of the files 

models file

~o
<STREAMINFO> 1 39
<VECSIZE> 39<NULLD><MFCC_D_A_0><DIAGC>
~s "silst"
<NUMMIXES> 6
<MIXTURE> 1 1.678619e-01
<MEAN> 39
.....
<VARIANCE> 39
...
.
.
~h "eight"
<BEGINHMM>
<NUMSTATES> 18
<STATE> 2
<NUMMIXES> 3
<MIXTURE> 1 3.110083e-01
<MEAN> 39
.....
.
.
<ENDHMM>
~h "one"
<BEGINHMM>
<NUMSTATES> 18
<STATE> 2
<NUMMIXES> 3
<MIXTURE> 1 2.197530e-01
<MEAN> 39
 -9.671743e+00 
.
.

macros file

~o
<STREAMINFO> 1 39
<VECSIZE> 39<NULLD><MFCC_D_A_0><DIAGC>
~v "varFloor1"
<VARIANCE> 39
 4.932017e-01 4.889575e-01 4.480567e-01....



my question is : how to import this two files to htk and use it to recognise digit ? what are the steps 

 

 

--- (Edited on 8/21/2017 10:07 am [GMT-0500] by yacine87) ---

--- (Edited on 8/21/2017 12:26 pm [GMT-0500] by yacine87) ---

PreviousNext