Sparse kernel cepstral coefficients (SKCC): Inner-product based features for noise-robust speech recognition.
Amin FazelShantanu ChakrabarttyPublished in: ISCAS (2011)
Keyphrases
- speech recognition
- cepstral coefficients
- noisy environments
- speech signal
- sparse kernel
- hidden markov models
- feature set
- language model
- speech recognition systems
- automatic speech recognition
- speaker identification
- pattern recognition
- feature extraction
- speaker verification
- feature vectors
- speaker recognition
- noise reduction
- audio signal
- computer vision
- neural network
- mel frequency cepstral coefficients
- signal to noise ratio
- edge detection