Bachelor Honours Project - Deep Speaker Embedding Across Languages

Mar 1, 2023ยท
Jiaying Fang
Jiaying Fang
ยท 0 min read
Speaker Verification Across Different Languages Adaptation Result
Abstract
We implemented deep speaker embedding for speaker verification with a domain loss to alleviate the languages mismatch problem. The performance of the ECAPA-TDNN (pre-trained using the English dataset) on the unlabelled Chinese dataset has improved by 10% with the MMD-based domain loss. This project has won the Honours Project - Technical Excellence Award.
Publication
Bachelor Thesis