Matrix quantization with vector quantization error compensation for robust speech recognition.
Lin CongSaf AsgharPublished in: MMSP (1998)
Keyphrases
- vector quantization
- speech recognition
- distortion measure
- error compensation
- speaker recognition
- image compression
- noisy environments
- successive approximation
- codebook design
- vector quantizer
- scalar quantization
- speech signal
- hidden markov models
- language model
- tree structured vector quantization
- pattern recognition
- automatic speech recognition
- speaker identification
- competitive learning
- vector quantized
- speech recognition systems
- compression ratio
- image coding
- multiresolution