Multi-resolution spectral input for convolutional neural network-based speech recognition.
László TóthPublished in: SpeD (2017)
Keyphrases
- speech recognition
- multiresolution
- cepstral coefficients
- hidden markov models
- speech synthesis
- pattern recognition
- language model
- speaker adaptation
- speech signal
- speech processing
- neural network
- automatic speech recognition
- speaker identification
- speech recognizer
- speech understanding
- speech recognition technology
- speech recognition systems
- speech recognition errors
- speech recognizers
- noisy environments
- speaker independent
- keyword spotting
- spectral analysis
- maximum likelihood
- speech retrieval
- computer vision
- information retrieval