%Aigaion2 BibTeX export from Idiap Publications %Saturday 21 December 2024 05:34:47 PM @INPROCEEDINGS{Madikeri_INTERSPEECH2018_2018, author = {Madikeri, Srikanth and Dey, Subhadeep and Motlicek, Petr}, keywords = {deep neural networks, i-vector, speaker recognition}, projects = {Idiap}, title = {Analysis of Language Dependent Front-End for Speaker Recognition}, booktitle = {Proceedings of Interspeech 2018}, volume = {1-6}, year = {2018}, pages = {1101-1105}, location = {Hyderabad, INDIA}, issn = {2308-457X}, isbn = {978-1-5108-7221-9}, doi = {10.21437/Interspeech.2018-2071}, abstract = {In Deep Neural Network (DNN) i-vector based speaker recognition systems, acoustic models trained for Automatic Speech Recognition are employed to estimate sufficient statistics for i-vector modeling. The DNN based acoustic model is typically trained on a wellresourced language like English. In evaluation conditions where enrollment and test data are not in English, as in the NIST SRE 2016 dataset, a DNN acoustic model generalizes poorly. In such conditions, a conventional Universal Background Model/Gaussian Mixture Model (UBM/GMM) based i-vector extractor performs better than the DNN based i-vector system. In this paper, we address the scenario in which one can develop a Automatic Speech Recognizer with limited resources for a language present in the evaluation condition, thus enabling the use of a DNN acoustic model instead of UBM/GMM. Experiments are performed on the Tagalog subset of the NIST SRE 2016 dataset assuming an open training condition. With a DNN i-vector system trained for Tagalog, a relative improvement of 12.1\% is obtained over a baseline system trained for English.} }