An Investigation on the Accuracy of Truncated DKLT Representation for Speaker Identification With Short Sequences of Speech Frames.
Giorgio BiagettiPaolo CrippaLaura FalaschettiSimone OrcioniClaudio TurchettiPublished in: IEEE Trans. Cybern. (2017)
Keyphrases
- speaker identification
- speech recognition
- speech signal
- speech processing
- speaker recognition
- gaussian mixture model
- hidden markov models
- broadcast news
- noisy environments
- speaker dependent
- feature extraction
- speaker independent
- image representation
- expectation maximization
- classification accuracy
- natural language
- information retrieval
- prediction accuracy
- em algorithm
- multi modal
- video frames
- maximum likelihood
- pattern recognition
- bayesian networks
- speech recognizer
- image processing
- computer vision