A comparative analysis between Conformer-Transducer, Whisper, and wav2vec2 for improving the child speech recognition

Research output: Chapter in Book or Conference Publication/ProceedingConference Publicationpeer-review

10 Citations (Scopus)

Abstract

Automatic Speech Recognition (ASR) systems have progressed significantly in their performance on adult speech data; however, transcribing child speech remains challenging due to the acoustic differences in the characteristics of child and adult voices. This work aims to explore the potential of adapting state-of-the-art Conformer-transducer models to child speech to improve child speech recognition performance. Furthermore, the results are compared with those of self-supervised wav2vec2 models and semi-supervised multi-domain Whisper models that were previously finetuned on the same data. We demonstrate that finetuning Conformer-transducer models on child speech yields significant improvements in ASR performance on child speech, compared to the non-finetuned models. We also show Whisper and wav2vec2 adaptation on different child speech datasets. Our detailed comparative analysis shows that wav2vec2 provides the most consistent performance improvements among the three methods studied.

Original languageEnglish
Title of host publication2023 International Conference on Speech Technology and Human-Computer Dialogue, SpeD 2023
EditorsDragos Burileanu
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages42-47
Number of pages6
ISBN (Electronic)9798350327977
DOIs
Publication statusPublished - 2023
Event12th International Conference on Speech Technology and Human-Computer Dialogue, SpeD 2023 - Bucharest, Romania
Duration: 25 Oct 202327 Oct 2023

Publication series

Name2023 International Conference on Speech Technology and Human-Computer Dialogue, SpeD 2023

Conference

Conference12th International Conference on Speech Technology and Human-Computer Dialogue, SpeD 2023
Country/TerritoryRomania
CityBucharest
Period25/10/2327/10/23

Keywords

  • Automatic Speech Recognition
  • Child Speech Recognition
  • CMU-Kids
  • Conformer-transducer
  • MyST
  • PF-STAR
  • wav2vec2
  • Whisper model

Fingerprint

Dive into the research topics of 'A comparative analysis between Conformer-Transducer, Whisper, and wav2vec2 for improving the child speech recognition'. Together they form a unique fingerprint.

Cite this