Is there any influence of the mismatch between the language used for training the system hyperparameters (TVS, LDA, and PLDA hyperparameters ) and the system users' language on the performance of the speaker recognition system ??
Thanks in advance .
Hi Ayoub,
Language mismatch between train and test data can surely affect the performance of the SR system (the error rate can be doubled or even worse if there is a mismatch between languages used to train the GMM, TV matrix and the PLDA model), I think that this paper can be a good starting point [1], it also proposes a phoneme histogram normalization technique to match the phonetic spaces of train and test languages. A possible solution is to use many languages to train your system, i.e. GMM, TV matrix and PLDA (using NIST SRE data) or other databases. Some systems [2] use up to 11 languages.
---
References :
[1] Abhinav Misra, John H. L. Hansen , Spoken language mismatch in speaker verification: An investigation with NIST-SRE and CRSS Bi-Ling corpora
[2] Pavel Matejka et al, Analysis of DNN approaches to speaker identification.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs