Source code used in the NIST SRE2016 challenge
In addition to the main bash script (run.sh), this repository contains some code augmentations specific to Kaldi:
NOTE: NDA Hasn't improved LDA results in our in-home experiments. May contain some bugs in the implementation. Feel free to use this project in any capacity on top of your kaldi-based SID systems. Please let me know if you have any comments, or corrections.
Dependencies:
Citations:
Probabilistic linear discriminant analysis for robust speaker identification in co-channel speech. In Sixteenth Annual Conference of the International Speech Communication Association.
Shokouhi, N., & Hansen, J. H. (2015).UTD-CRSS systems for 2016 nist speaker recognition evaluation. arXiv preprint arXiv:1610.07651.
Zhang, C., Bahmaninezhad, F., Ranjan, S., Yu, C., Shokouhi, N., & Hansen, J. H. (2016).Navid Shokouhi
2016