Low-latency speaker diarization based on Bayesian information criterion with multiple phoneme classes.
Takahiro OkuShoei SatoAkio KobayashiShinichi HommaToru ImaiPublished in: ICASSP (2012)
Keyphrases
- speaker diarization
- bayesian information criterion
- low latency
- speech recognition
- model selection
- high throughput
- broadcast news
- high speed
- mixture model
- gaussian mixture model
- highly efficient
- automatic speech recognition
- real time
- virtual machine
- image segmentation
- information retrieval
- dimensionality reduction
- stream processing
- hidden markov models
- computer vision