Exploiting Speaker Embeddings for Improved Microphone Clustering and Speech Separation in ad-hoc Microphone Arrays.
Stijn KindtJenthe ThienpondtNilesh MadhuPublished in: ICASSP (2023)
Keyphrases
- automatic speech recognition
- speaker diarization
- speech recognition
- bayesian information criterion
- sound source
- speech signal
- audio stream
- broadcast news
- audio visual
- clustering algorithm
- hidden markov models
- clustering method
- hierarchical clustering
- speaker verification
- cluster analysis
- k means
- spontaneous speech
- speaker recognition
- acoustic features
- categorical data
- multi modal
- low dimensional
- speech segments
- noisy environments
- fuzzy clustering
- audio features
- unsupervised learning
- data clustering
- vector space
- speech sounds
- mel frequency cepstral coefficients
- speaker identification
- spectral clustering
- visual information
- gaussian mixture model
- mixture model
- language model
- pattern recognition