End-to-End Integration of Speech Separation and Voice Activity Detection for Low-Latency Diarization of Telephone Conversations.
Giovanni MorroneSamuele CornellLuca SerafiniEnrico ZovatoAlessio BruttiStefano SquartiniPublished in: CoRR (2023)
Keyphrases
- end to end
- voice activity detection
- low latency
- high bandwidth
- noisy environments
- speech recognition
- high speed
- high throughput
- real time
- speaker diarization
- stream processing
- highly efficient
- multipath
- virtual machine
- ad hoc networks
- speech signal
- congestion control
- pattern recognition
- noise reduction
- gaussian mixture model
- hidden markov models
- multimedia