Non-stationary feature extraction for automatic speech recognition.
Zoltán TüskePavel GolikRalf SchlüterFriedhelm R. DrepperPublished in: ICASSP (2011)
Keyphrases
- non stationary
- automatic speech recognition
- speech signal
- feature extraction
- speech recognition
- speaker identification
- word error rate
- broadcast news
- hidden markov models
- noisy environments
- conversational speech
- spoken words
- spontaneous speech
- feature vectors
- speech retrieval
- feature selection
- autoregressive
- vocal tract
- word recognition
- adaptive algorithms
- pattern recognition
- empirical mode decomposition
- machine learning
- image processing
- face recognition
- image classification
- acoustic features
- feature space
- extracted features
- recognition errors
- data mining
- speech sounds
- speech corpus
- multiresolution
- feature set
- frequency domain