Recognition of Noisy Speech by Starting the Likelihood Calculation at Voiced Segments.
Hans-Günter HirschFrank KremerPublished in: ITG Symposium on Speech Communication (2014)
Keyphrases
- noisy speech
- speech signal
- speech recognition
- automatic speech recognition systems
- hidden markov models
- speech enhancement
- automatic speech recognition
- object recognition
- background noise
- noisy environments
- maximum likelihood
- non stationary
- machine learning
- signal to noise ratio
- noise reduction
- speech segments
- additive noise
- speaker identification
- feature extraction
- pattern recognition
- speaker recognition
- neural network