Improved CS-CELP speech coding in a noisy environment using a trained sparse conjugate codebook.
Akitoshi KataokaSachiko HosakaJotaro IkedoTakehiro MoriyaShinji HayashiPublished in: ICASSP (1995)
Keyphrases
- noisy environments
- speech signal
- linear prediction
- speech recognition
- soft assignment
- speech enhancement
- speaker verification
- noise reduction
- automatic speech recognition
- training set
- speaker identification
- background noise
- linear predictive
- voice activity detection
- speech quality
- noisy speech
- vector quantization
- non stationary
- computer vision
- hidden markov models
- multilayer perceptron
- compressive sampling
- image processing
- feature vectors
- compressive sensing
- bag of words
- coding scheme
- speaker recognition
- vocal tract
- image compression
- sparse representation
- machine learning