This repository provides the implementation of an advanced deep embedding learning architecture for Text-Dependent Speaker Verification (TD-SV), originally presented at INTERSPEECH 2020. Our approach focuses on capturing both short-term acoustic details and long-term temporal contexts to excel in challenging far-field environments.

If you find this code useful for your research, please cite our paper:

@inproceedings{zhang2020deep,
  title={Deep Embedding Learning for Text-Dependent Speaker Verification.},
  author={Zhang, Peng and Hu, Peng and Zhang, Xueliang},
  booktitle={INTERSPEECH},
  pages={3461--3465},
  year={2020}
}