Preventing an external acoustic noise from being misrecognized as a speech recognition object by confirming the lip movement image signal.
Soo-Jong LeeJun ParkEung-Kyeu KimPublished in: INTERSPEECH (2007)
Keyphrases
- speech recognition
- noisy speech
- background noise
- noisy environments
- additive noise
- speech signal
- speech recognition systems
- low frequency
- spectral subtraction
- hidden markov models
- language model
- speech synthesis
- speech recognizers
- speech processing
- speaker independent
- pattern recognition
- automatic speech recognition
- power spectrum
- high frequency
- speaker identification
- wiener filter
- acoustic features
- speech recognizer
- signal to noise ratio
- image classification
- multiscale
- image segmentation
- low contrast
- speech recognition technology
- edge detection
- low level
- speaker verification
- image retrieval
- visual speech
- audio visual speech recognition
- image restoration
- signal processing
- gaussian noise
- noise reduction
- frequency domain
- non stationary
- machine learning
- neural network