Wavesplit: End-to-End Speech Separation by Speaker Clustering.
Neil ZeghidourDavid GrangierPublished in: IEEE ACM Trans. Audio Speech Lang. Process. (2021)
Keyphrases
- end to end
- speaker clustering
- speaker identification
- feature transformation
- gaussian mixture model
- discriminant analysis
- partially supervised
- dimensionality reduction
- speech recognition
- speech signal
- congestion control
- cosine distance
- noisy environments
- audio visual
- automatic speech recognition
- pattern recognition
- mixture model
- broadcast news
- scale invariant
- principal components
- euclidean distance
- acoustic features
- em algorithm
- image quality
- feature vectors