Speaker recognition for Timit dataset

[Tushar nimbhorkar]

Hi 

I am working on a project of speaker recognition on Timit Dataset. I want to extract ivectors from the dataset. So that i can feed them in my model.

Any Tutorial regarding it. I want my feature vector for neural network(Theano). 

TIA

[remi....@gmail.com]

You can have a look at the swbd recipe. In the nnet3 section, you can find run_ivector_common.sh, this trains an ivector extractor and then extract the ivectors on the data. You will have to adapt this script to your data.

[alyam...@gmail.com]

Hi, do you have TIMIT database? I also work on speker recognition project 

четверг, 10 марта 2016 г., 14:20:57 UTC+6 пользователь Tushar nimbhorkar написал:

[Tushar nimbhorkar]

Yes i have timit dataset.

What algorithm/process you working on?

[alyam...@gmail.com]

I work on LSTM based algorithm using keras.

четверг, 17 марта 2016 г., 14:56:10 UTC+6 пользователь Tushar nimbhorkar написал:

[David Snyder]

It might be helpful if you look at the run.sh scripts in the examples egs/sre10/v1 and egs/sre08/v1. These recipes demonstrate standard ivector systems with GMM-UBMs on a speaker recognition task. They don't use Timit, but it should give you an idea on how to set it up.