Online meeting recognition in noisy environments with time-frequency mask based MVDR beamforming.
Shoko ArakiNobutaka ItoMarc DelcroixAtsunori OgawaKeisuke KinoshitaTakuya HiguchiTakuya YoshiokaDung T. TranShigeki KaritaTomohiro NakataniPublished in: HSCMA (2017)
Keyphrases
- noisy environments
- speech enhancement
- noise reduction
- digit recognition
- noisy speech
- speech recognition
- word recognition
- single channel
- background noise
- frequency domain
- speech signal
- speaker verification
- speaker identification
- signal processing
- recognition rate
- signal to noise ratio
- object recognition
- feature extraction
- multi channel
- linear prediction
- automatic speech recognition
- non stationary
- edge detection
- wavelet transform
- multiresolution
- pattern recognition
- bayesian networks
- machine learning
- voice activity detection