The Munich 2011 CHiME Challenge Contribution: NMF-BLSTM Speech Enhancement and Recognition for Reverberated Multisource Environments
Year: 2011
Type: article
Abstract: We present the Munich contribution to the PASCAL 'CHiME' Speech Separation and Recognition Challenge: Our approach combines source separation by supervised convolutive non-negative matrix factorisation (NMF) with our tandem recogniser that augments acoustic features by word predictions of a Long Short-Term Memory recurrent neural network in a multi-stream Hidden Markov Model.The performance of our source separation approach is demonstrated in a sequence of gradually refined speech recognisers.While NMF drastically improves performance for all investigated recognisers, best results are obtained with the multi-stream approach along with a novel adaptation technique for noise dictionaries in supervised NMF.On the final Challenge test set, the proposed system delivers an average keyword recognition accuracy of 87.86 % across SNRs ranging from -6 to 9 dB, reducing the error rate from 44 % to 12 % compared to the Challenge baseline.
Institution: Technical University of Munich
Language: English
Topic: Speech and Audio Processing
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2336741987
