UIAI System for Short-Duration Speaker Verification Challenge 2020
share
In this work, we present the system description of the UIAI entry for the short-duration speaker verification (SdSV) challenge 2020. Our focus is on Task 1 dedicated to text-dependent speaker verification. We investigate different feature extraction and modeling approaches for automatic speaker verification (ASV) and utterance verification (UV). We have also studied different fusion strategies for combining UV and ASV modules. Our primary submission to the challenge is the fusion of seven subsystems which yields a normalized minimum detection cost function (minDCF) of 0.072 and an equal error rate (EER) of 2.14 identification based model with phone-discriminative bottleneck features gives a normalized minDCF of 0.118 and achieves 19 state-of-the-art challenge baseline.
READ FULL TEXT
