Spectral-temporal receptive fields and MFCC balanced feature extraction for noisy speech recognition.
Jia-Ching WangChang-Hong LinEn-Ting ChenPao-Chi ChangPublished in: APSIPA (2014)
Keyphrases
- speech recognition
- receptive fields
- cepstral coefficients
- noisy environments
- feature extraction
- biologically inspired
- space time
- speech signal
- natural images
- speaker identification
- automatic speech recognition
- pattern recognition
- hidden markov models
- image representation
- language model
- mel frequency cepstral coefficients
- spatio temporal
- saliency map
- speaker verification
- speech recognizer
- face recognition
- speech recognition systems
- hidden layer
- input space
- image classification
- dimensionality reduction
- feature selection
- linear prediction
- audio signal
- feature set
- feature vectors
- multiscale
- bayesian networks
- multi modal
- higher order
- principal component analysis