Comparing noise compensation methods for robust prediction of acoustic speech features from MFCC vectors in noise.
Ben P. MilnerJonathan DarchIbrahim AlmajaiSaeed VaseghiPublished in: EUSIPCO (2008)
Keyphrases
- greater robustness
- speech recognition
- feature vectors
- noisy environments
- image noise
- feature extraction
- noisy speech
- noise sensitivity
- speech signal
- feature descriptors
- noise reduction
- noisy data
- noise level
- hybrid method
- prediction accuracy
- speech recognition systems
- speech enhancement
- prediction error
- discriminatory power
- signal to noise ratio
- low level
- feature space