VoxForge
VoxForge
Hello:
I am trying to build a Command System based on Automatic Voice Recognition. The program will take Albanian voice input and execute commands.
I will be training it with samples from 5-10 people. I have chosen a combination of HTK and Julius. I did also follow the tutorial on how to create an acoustic model:
http://www.voxforge.org/home/dev/acousticmodels/linux/create/htkjulius/how-to
I was wondering if any of you could give me any advices on how to get started or whether I should be using word or phoneme recognition.
Thank you in advance!
--- (Edited on 7/9/2010 11:23 pm [GMT-0500] by ) ---
Hi! "how to get started" Maybe you want to take a look at my Albanian pronunciation dictionary? It contains 576613 Albanian words. Greetings, Ralf
--- (Edited on 2010-07-10 8:12 pm [GMT-0500] by ralfherzog) ---
>I am trying to build a Command System based on Automatic Voice
>Recognition.
see here: What is a Dialog Manager? for different approaches to creating a dialog manager.
>how to get started or whether I should be using word or phoneme
>recognition.
from this post: Whole-Word Recognition
[...] the question is whether to use whole word HMMs or phonetic unit (phoneme) HMMs. Whole word HMM are usefull if you want to do IWR [Isolated Word Recognition] and do not have enough training data. Otherwise, just use phoneme HMM (i.e. the standard described in all tutorials).
--- (Edited on 7/14/2010 4:35 pm [GMT-0400] by kmaclean) ---