Hybrid input spaces for exemplar-based noise robust speech recognition using coupled dictionaries.
Deepak BabyHugo Van hammePublished in: EUSIPCO (2015)
Keyphrases
- noisy environments
- image noise
- input data
- speech recognition
- salt pepper
- geometric distortions
- computationally efficient
- missing data
- signal to noise ratio
- noise reduction
- estimation error
- imaging artifacts
- greater robustness
- noise model
- speech signal
- noisy data
- non stationary
- data sets
- automatic speech recognition
- additive noise
- sparse representation
- natural language processing
- background noise
- spoken language
- speaker identification
- text input
- information extraction
- pattern recognition
- image processing