Publication Details
A multi purpose and large scale speech corpus in Persian and English for speaker and speech Recognition: the DeepMine database
Černocký Jan, prof. Dr. Ing. (DCGM)
Burget Lukáš, doc. Ing., Ph.D. (DCGM)
speech database, text-dependent, text-independent, speaker verification, speech
recognition
DeepMine is a speech database in Persian and English designed to build and
evaluate text-dependent, text-prompted, and textindependent speaker verification,
as well as Persian speech recognition systems. It contains more than 1850
speakers and 540 thousand recordings overall, more than 480 hours of speech are
transcribed. It is the first public large-scale speaker verification database in
Persian, the largest public text-dependent and text-prompted speaker verification
database in English, and the largest public evaluation dataset for
text-independent speaker verification. It has a good coverage of age, gender, and
accents. We provide several evaluation protocols for each part of the database to
allow for research on different aspects of speaker verification. We also provide
the results of several experiments that can be considered as baselines: HMM-based
i-vectors for text-dependent speaker verification, and HMM-based as well as
state-of-the-art deep neural network based ASR. We demonstrate that the database
can serve for training robust ASR models.
@inproceedings{BUT161477,
author="Hossein {Zeinali} and Jan {Černocký} and Lukáš {Burget}",
title="A multi purpose and large scale speech corpus in Persian and English for speaker and speech Recognition: the DeepMine database",
booktitle="IEEE Automatic Speech Recognition and Understanding Workshop - Proceedings (ASRU)",
year="2019",
pages="397--402",
publisher="IEEE Signal Processing Society",
address="Sentosa, Singapore",
doi="10.1109/ASRU46091.2019.9003882",
isbn="978-1-7281-0306-8",
url="https://www.fit.vut.cz/research/publication/12153/"
}