| Document Date: 2014-03-11 09:25:13
Open Document File Size: 88,39 KBShare Result on Facebook
City Kyoto / Cambridge / Lincoln / / Company FBANK / NTT Communication Science Laboratories / Hybrid/FBANK / HMM Tandem/MFCC Hybrid/MFCC Hybrid/FBANK Hybrid/FBANK / Toshiba Research Europe Ltd / / Country Japan / / / Event Reorganization / / IndustryTerm speech recognition systems / employed confusion network / log mel-filter bank channel outputs / tandem systems / deep belief networks / neural network / log mel-filter bank features / large scale in-car speech recognition task / context-depencent deep neural networks / dereverberation processing / front-end processing / dereverberation algorithm / dereverberation processing offers / neural networks / ubiquitous recording devices / automatic speech recognition systems / microphone array processing / less sophisticated systems / noise reduction algorithm / front-end processor / web data / front-end processing approach / deep neural networks / mel-filter bank outputs / / MusicAlbum SPEECH / / Organization Cambridge Research Lab. / Mark J. F. Gales† †Cambridge University Engineering Department / / Person J. Kadlec / V / Mark J. F. Gales / / Position online speaker / speaker / applying speaker / / Product Sangean Table Top Portable Audio Device / / PublishedMedium Machine Learning / / RadioStation DNN / / Technology speech recognition / neural network / cepstral-domain MMSE noise reduction algorithm / smartphones / beamforming algorithm / front-end processor / Machine Learning / STFT-domain dereverberation algorithm / BeamformIt algorithm / speech recognition system / / URL www.idiap.ch/dataset/ami / /
SocialTag |
IMPACT OF SINGLE-MICROPHONE DEREVERBERATION ON DNN-BASED MEETING TRANSCRIPTION SYSTEMS Takuya Yoshioka†,‡ , Xie Chen† , Mark J. F. Gales† †Cambridge University Engineering Department, Cambridge, UK ‡NTT Commu