Deep neural networks for automatic speaker recognition do not learn supra-segmental temporal features.
Daniel NeururerVolker DellwoThilo StadelmannPublished in: Pattern Recognit. Lett. (2024)
Keyphrases
- speaker recognition
- neural network
- gaussian mixture model
- vector quantization
- speaker verification
- probabilistic neural network
- speaker identification
- pattern recognition
- speech recognition
- artificial neural networks
- multi modal
- image compression
- neural network model
- multilayer perceptron
- hidden markov models
- genetic algorithm
- self organizing maps
- nearest neighbor
- speech signal
- fuzzy logic