EEND-DEMUX: End-to-End Neural Speaker Diarization via Demultiplexed Speaker Embeddings.
Sung Hwan MunMin Hyun HanCanyeong MoonNam Soo KimPublished in: CoRR (2023)
Keyphrases
- speaker diarization
- end to end
- network architecture
- speech recognition
- neural network
- audio stream
- bayesian information criterion
- admission control
- congestion control
- broadcast news
- dimensionality reduction
- machine learning
- scalable video
- speaker verification
- pattern recognition
- non stationary
- visual information
- multi modal
- video sequences