Feature compensation based on the normalization of vocal tract length for the improvement of emotion-affected speech recognition.
Masoud GeravanchizadehElnaz ForouhandehMeysam BashirpourPublished in: EURASIP J. Audio Speech Music. Process. (2021)
Keyphrases
- speech recognition
- vocal tract
- speech synthesis
- length normalization
- speech signal
- passage retrieval
- language model
- automatic speech recognition
- hidden markov models
- cepstral coefficients
- pattern recognition
- speech recognizer
- speaker identification
- speech enhancement
- document retrieval
- speech recognition systems
- text to speech
- background noise
- linear prediction
- machine learning
- noisy environments
- speaker adaptation
- multiscale
- spectral analysis
- n gram