ACL RD-TEC 1.0 Summarization of H92-1055
Paper Title:
MULTIPLE APPROACHES TO ROBUST SPEECH RECOGNITION
MULTIPLE APPROACHES TO ROBUST SPEECH RECOGNITION
Authors: Richard M. Stern and Fu-Hua Liu and Yoshiaki Ohshima and Thomas M. Sullivan and Alejandro Acero
Primarily assigned technology terms:
- algorithm
- array microphone
- atis
- auditory processing
- automatic speech recognition
- blind deconvolution
- cdcn algorithm
- cepstral normalization
- database
- deconvolution
- error analysis
- microphone array
- modelling
- nonlinear warping
- normalization
- pre-processing
- processing
- recognition
- recognition system
- recognition systems
- robust speech recognition
- sdcn
- signal processing
- snr-dependent cepstral normalization
- speaker-independent recognition
- spectral normalization
- spectral subtraction
- speech recognition
- speech recognition system
- speech recognition systems
- speech system
- sphinx-ii
- spoken language systems
- tuning
Other assigned terms:
- additive noise
- approach
- array
- auditory model
- beam
- benchmark
- case
- cepstral domain
- computational complexity
- continuous-speech
- document
- error rate
- evaluations
- events
- fact
- french
- implementation
- knowledge
- method
- noise
- perplexity
- priori
- procedure
- process
- recognition accuracy
- segments
- sentences
- signal
- silence
- speech signal
- spoken language
- temporal information
- test set
- training
- training data
- training database
- understanding
- utterance
- white noise
