Blind speaker counting in highly reverberant environments by clustering coherence features.
Shahab PashaJacob DonleyChristian H. RitzPublished in: APSIPA (2017)
Keyphrases
- feature extraction
- additional features
- image features
- real world
- unsupervised learning
- mel frequency cepstral coefficients
- speech signal
- dynamic environments
- feature set
- feature vectors
- feature space
- clustering algorithm
- low level
- co occurrence
- object recognition
- false positives
- spatial information
- high dimensionality
- data clustering
- similarity function
- fuzzy clustering
- video sequences
- speaker identification
- machine learning