Synthesis Speech Based Data Augmentation for Low Resource Children ASR

Virender Kadyan, Hemant Kathania*, Prajjval Govil, Mikko Kurimo

*Tämän työn vastaava kirjoittaja

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussaConference contributionScientificvertaisarvioitu

5 Lataukset (Pure)

Abstrakti

Successful speech recognition for children requires large training data with sufficient speaker variability. The collection of such a training database of children’s voices is challenging and very expensive for zero/low resource language like Punjabi. In this paper, the data scarcity issue of the low resourced language Punjabi is addressed through two levels of augmentation. The original training corpus is first augmented by modifying the prosody parameters for pitch and speaking rate. Our results show that the augmentation improves the system performance over the baseline system. Then the augmented data combined with original data and used to train the TTS system to generate synthesis data and extended dataset is further used for augmented by generating children’s utterances using text-to-speech synthesis and sampling the language model with methods that increase the acoustic and lexical diversity. The final speech recognition performance indicates a relative improvement of 50.10% with acoustic and 57.40% with language diversity based augmentation in comparison to that of the baseline system respectively.

AlkuperäiskieliEnglanti
OtsikkoSpeech and Computer - 23rd International Conference, SPECOM 2021, Proceedings
ToimittajatAlexey Karpov, Rodmonga Potapova
Sivut317-326
Sivumäärä10
DOI - pysyväislinkit
TilaJulkaistu - 2021
OKM-julkaisutyyppiA4 Artikkeli konferenssijulkaisuussa
TapahtumaInternational Conference on Speech and Computer - Virtual, Online
Kesto: 27 syyskuuta 202130 syyskuuta 2021
Konferenssinumero: 23

Julkaisusarja

NimiLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Vuosikerta12997 LNAI
ISSN (painettu)0302-9743
ISSN (elektroninen)1611-3349

Conference

ConferenceInternational Conference on Speech and Computer
LyhennettäSPECOM
KaupunkiVirtual, Online
Ajanjakso27/09/202130/09/2021

Sormenjälki

Sukella tutkimusaiheisiin 'Synthesis Speech Based Data Augmentation for Low Resource Children ASR'. Ne muodostavat yhdessä ainutlaatuisen sormenjäljen.

Siteeraa tätä