Combination of distortion-robust feature extraction and neural noise reduction for ASR.
Michael TrompfRalf RichterHarald EckhardtHeidi HackbarthPublished in: EUROSPEECH (1993)
Keyphrases
- noise reduction
- noisy environments
- feature extraction
- edge preserving
- signal to noise ratio
- edge detection
- noise level
- speech enhancement
- noise removal
- impulse noise
- speech recognition
- automatic speech recognition
- median filter
- edge enhancement
- preprocessing
- geometric distortions
- noise detection
- feature selection
- hearing aids
- texture features
- wavelet transform
- multiresolution
- robust estimation
- frequency domain
- principal component analysis
- hidden markov models
- edge preservation
- pattern recognition
- face recognition