The second &amp;#x2018;chime&amp;#x2019; speech separation and recognition challenge: Datasets, tasks and baselines

Vincent, Emmanuel; Barker, Jon; Watanabe, Shinji; Roux, Jonathan Le; Nesta, Francesco; Matassoni, Marco

doi:10.1109/icassp.2013.6637622

Cited by 230 publications

(140 citation statements)

References 10 publications

Supporting

Mentioning

138

Contrasting

Unclassified

Order By: Relevance

“…The effectiveness of the proposed framework is evaluated on ChiME-WSJ0 corpus [15] using MFCC features. A DDA is first trained on the training set.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Speech feature denoising and dereverberation via deep autoencoders for noisy reverberant speech recognition

Feng¹,

Zhang²,

Glass³

2014

2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

197

View full text Add to dashboard Cite

Denoising autoencoders (DAs) have shown success in generating robust features for images, but there has been limited work in applying DAs for speech. In this paper we present a deep denoising autoencoder (DDA) framework that can produce robust speech features for noisy reverberant speech recognition. The DDA is first pre-trained as restricted Boltzmann machines (RBMs) in an unsupervised fashion. Then it is unrolled to autoencoders, and fine-tuned by corresponding clean speech features to learn a nonlinear mapping from noisy to clean features. Acoustic models are re-trained using the reconstructed features from the DDA, and speech recognition is performed. The proposed approach is evaluated on the CHiME-WSJ0 corpus, and shows a 16-25% absolute improvement on the recognition accuracy under various SNRs.Index Terms-robust speech recognition, feature denoising, denoising autoencoder, deep neural network

show abstract

“…The effectiveness of the proposed framework is evaluated on ChiME-WSJ0 corpus [15] using MFCC features. A DDA is first trained on the training set.…”

Section: Discussionmentioning

confidence: 99%

“…We validate the effectiveness of our proposed DDA frontend denoising approach on track 2 of the second CHiME challenge [15]. Track 2 is a 5k medium-vocabulary speech recognition task in reverberant and noisy environment, whose utterances are taken from the Wall Street Journal database (WSJ0).…”

Section: Introductionmentioning

confidence: 99%

Speech feature denoising and dereverberation via deep autoencoders for noisy reverberant speech recognition

Feng¹,

Zhang²,

Glass³

2014

2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

197

View full text Add to dashboard Cite

show abstract

“…The small vocabulary track of the 2 nd CHiME Challenge (Vincent et al, 2013) addresses the problem of recognizing commands in a noisy living room. The clean utterances in the CHIME-2 data are taken from the GRID corpus (Cooke et al, 2006) which contains utterances from 34 speakers reading 6-word sequences of the form command-color-preposition-letterdigit-adverb.…”

Section: Chime-2mentioning

confidence: 99%

Noise robust exemplar matching with alpha–beta divergence

Yılmaz

Gemmeke

hamme

2016

Speech Communication

View full text Add to dashboard Cite

The noise robust exemplar matching (N-REM) framework performs automatic speech recognition using exemplars, which are the labeled spectrographic representations of speech segments extracted from training data. By incorporating a sparse representations formulation, this technique remedies the inherent noise modeling problem of conventional exemplar matching-based automatic speech recognition systems. In this framework, noisy speech segments are approximated as a sparse linear combination of the exemplars of multiple lengths, each associated with a single speech unit such as words, half-words or phones. On account of the reconstruction error-based back end, the recognition accuracy highly depends on the congruence of the speech features and the divergence metric used to compare the speech segments with exemplars. In this work, we replace the conventional KullbackLeibler divergence (KLD) with a generalized divergence family called the Alpha-Beta divergence with two parameters, α and β, in conjunction with mel-scaled magnitude spectral features. The proposed recognizer traverses the (α,β) plane depending on the amount of contamination to provide better separation of speech and noise sources. Moreover, we apply our recently proposed active noise exemplar selection (ANES) technique in a more realistic scenario where the target utterances are degraded by genuine room noise. Recognition experiments on the small vocabulary track of the 2 nd CHiME Challenge and the AURORA-2 database have shown that the novel recognizer with the AB divergence and ANES outperforms the baseline system using the generalized KLD with tuned sparsity, especially at lower SNR levels.

show abstract

“…Development of such processing methods is especially important for the applications like "smart home" when microphones are located in several parts of the room and both location and orientation of speaker's head are unknown in advance. In order to promote the development of multi-microphone approaches in robust speech recognition, several international competitions like CHiME (Computational Hearing in Multisource Environments) Challenge [6][7][8][9] have been organized in recent years.…”

Section: State Of the Art In Study Areamentioning

confidence: 99%

Investigation and Development of Methods for Improving Robustness of Automatic Speech Recognition Algorithms in Complex Acoustic Environments

Korenevsky

Matveev

Yakovlev

2017

Proceedings of the Scientific-Practical Conference "Research and Development - 2016"

View full text Add to dashboard Cite

Aims and objectives of the study are described; state-of-the-art techniques in the study area are outlined. Several effective approaches proposed in the study and targeted at robustness improvement in complex acoustic environments are described. They are multichannel alignment algorithm, vector Taylor series-based features compensation with phase-term modeling, and environment adaptation method based on GMM-derived features. Experimental results analysis and comparison to state of the art are presented.

show abstract

The second ‘chime’ speech separation and recognition challenge: Datasets, tasks and baselines

Cited by 230 publications

References 10 publications

Speech feature denoising and dereverberation via deep autoencoders for noisy reverberant speech recognition

Speech feature denoising and dereverberation via deep autoencoders for noisy reverberant speech recognition

Noise robust exemplar matching with alpha–beta divergence

Investigation and Development of Methods for Improving Robustness of Automatic Speech Recognition Algorithms in Complex Acoustic Environments

Contact Info

Product

Resources

About