A Semi-Automatic Approach to Create Large Gender- and Age-Balanced Speaker Corpora: Usefulness of Speaker Diarization & Identification.
Rémi UroDavid DoukhanAlbert RilliardLaetitia LarcherAnissa-Claire AdgharouamaneMarie TahonAntoine LaurentPublished in: CoRR (2024)
Keyphrases
- semi automatic
- speaker diarization
- fully automatic
- speech recognition
- domain ontology
- semi automatically
- speaker verification
- gold standard
- broadcast news
- age groups
- bayesian information criterion
- natural language processing
- audio stream
- labor intensive
- speaker identification
- ontology mapping
- computational linguistics