Catalog-based single-channel speech-music separation for automatic speech recognition.
Cemil DemirMehmet Ugur DoganAli Taylan CemgilMurat SaraclarPublished in: SIU (2012)
Keyphrases
- single channel
- automatic speech recognition
- speech signal
- sound source
- audio features
- speech recognition
- acoustic features
- speech enhancement
- multi channel
- broadcast news
- word error rate
- speaker identification
- noisy environments
- hidden markov models
- vocal tract
- non stationary
- linear prediction
- spoken words
- additive noise
- recognition errors
- conversational speech
- speech corpus
- spontaneous speech
- speech retrieval
- music information retrieval
- speech synthesis
- independent component analysis
- frequency domain
- multi modal
- speech sounds
- wiener filter
- neural network
- image compression
- language model
- pattern recognition
- video sequences
- feature extraction
- computer vision