Speaker identification using hybrid Karhunen-Loeve transform and Gaussian mixture model approach.
Chih-Chien Thomas ChenChin-Ta ChenCheng-Kuan HouPublished in: Pattern Recognit. (2004)
Keyphrases
- speaker identification
- gaussian mixture model
- karhunen loeve transform
- mixture model
- speaker recognition
- image compression
- em algorithm
- feature vectors
- maximum likelihood
- expectation maximization
- feature transformation
- feature space
- unsupervised learning
- visual words
- feature extraction
- supervised learning
- image segmentation
- high dimensional
- speech signal
- pattern recognition
- bayesian information criterion