Speech enhancement using beamforming and non negative matrix factorization for robust speech recognition in the CHiME-3 challenge

Vu, Thanh Tung; Bigot, Benjamin; Chng, Eng Siong

doi:10.1109/asru.2015.7404826

Cited by 9 publications

(15 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Augmenting the training set by using individual microphone channels or performing semi-supervised adaptation on the test data also yielded consistent improvements on real and simulated data (Yoshioka et al, 2015). By contrast with these results, Vu et al (2015) found that, in the case when MVDR beamforming is applied, training the ASR backend on real data only improves the WER on real test data compared to training on real and simulated data. This confirms that the difference in the characteristics of enhanced real vs. simulated signals induced by MVDR carries over to the ASR backend too.…”

Section: Acoustic Modelingcontrasting

confidence: 48%

“…As expected, Bagchi et al (2015) and Fujita et al (2015) reported similar performance for these two techniques on real and simulated data. Single-channel enhancement based on nonnegative matrix factorization (NMF) of the power spectra of speech and noise has also been used and resulted in minor improvement on both real and simulated data (Bagchi et al, 2015;Vu et al, 2015).…”

Section: Source Separationmentioning

confidence: 99%

“…The performance improvements brought by these increasingly complex architectures appear to be consistent on real and simulated data. Table 8: WER (%) achieved on noisy data using various acoustic models trained on noisy real and simulated data (all channels) without sMBR (Yoshioka et al, 2015 It must be noted that, with the exception of Vu et al (2015), all challenge entrants trained GMM-HMM and DNN-HMM acoustic models on real and simulated data. Heymann et al (2015) found that discarding real data and training a GMM-HMM acoustic model on simulated data only increases the WER by 3% and 4% relative on real development and test data, respectively.…”

Section: Acoustic Modelingmentioning

confidence: 99%

See 2 more Smart Citations

An analysis of environment, microphone and data simulation mismatches in robust speech recognition

Vincent

Watanabe

Nugraha

et al. 2017

Computer Speech & Language

291

177

View full text Add to dashboard Cite

Section: Acoustic Modelingcontrasting

confidence: 48%

Section: Source Separationmentioning

confidence: 99%

Section: Acoustic Modelingmentioning

confidence: 99%

See 1 more Smart Citation

An analysis of environment, microphone and data simulation mismatches in robust speech recognition

Vincent

Watanabe

Nugraha

et al. 2017

Computer Speech & Language

291

177

View full text Add to dashboard Cite

“…Vu et al (2015) and Baby et al (2015) employ non-negative matrix factorisation approaches that exploit spectral diversity of speech and noise. Bagchi et al (2015) employ a DNN-based denoising autoencoder and El-Desoky Mousa et al (2015) perform feature denoising using a bidirectional Long Short-Term Memory (BLSTM).…”

Section: Target Enhancementmentioning

confidence: 99%

“…The simplest approach has been to apply utterance-based feature mean and variance normalization (Zhao et al, 2015;Fujita et al, 2015;Du et al, 2015;Wang et al, 2015). However, the two most effective techniques are transforming the DNN features using feature-space maximum likelihood linear regression (fMLLR) (Hori et al, 2015;Moritz et al, 2015;Vu et al, 2015;Sivasankaran et al, 2015;Tran et al, unpublished) or augmentation of the DNN features using either i-vectors, (e.g., Moritz et al, 2015;Zhuang et al, 2015), pitch-based features (Ma et al, 2015;Wang et al, 2015;Du et al, 2015) or bottleneck features (Tachioka et al, 2015), i.e., extracted from bottleneck layers in speaker classification DNNs. Where i-vectors have been used they may be either per-speaker (e.g., Prudnikov et al, 2015) or per-speaker-environment, (e.g.…”

Section: Feature Designmentioning

confidence: 99%

The third ‘CHiME’ speech separation and recognition challenge: Analysis and outcomes

Barker

Marxer

Vincent

et al. 2017

Computer Speech & Language

101

View full text Add to dashboard Cite

This paper presents the design and outcomes of the CHiME-3 challenge, the first open speech recognition evaluation designed to target the increasingly relevant multichannel, mobile-device speech recognition scenario. The paper serves two purposes. First, it provides a definitive reference for the challenge, including full descriptions of the task design, data capture and baseline systems along with a description and evaluation of the 26 systems that were submitted. The best systems re-engineered every stage of the baseline resulting in reductions in word error rate from 33.4% to as low as 5.8%. By comparing across systems, techniques that are essential for strong performance are identified. Second, the paper considers the problem of drawing conclusions from evaluations that use speech directly recorded in noisy environments. The degree of challenge presented by the resulting material is hard to control and hard to fully characterise. We attempt to dissect the various 'axes of difficulty' by correlating various estimated signal properties with typical system performance on a per session and per utterance basis. We find strong evidence of a dependence on signal-to-noise ratio and channel quality. Systems are less sensitive to variations in the degree of speaker motion. The paper concludes by discussing the outcomes of CHiME-3 in relation to the design of future mobile speech recognition evaluations.

show abstract

Tactile Sensing, Information, and Feedback via Wave Propagation

Shao

2022

Springer Series on Touch and Haptic Systems

View full text Add to dashboard Cite

Through my entire PhD study, I learned a lot from my colleagues in the RE Touch Lab. Discussions with Gregory Reardon about technical problems, like in physics or statistics, are always constructive and inspiring. Shantonu Biswas and Thanh Nho Do influenced me with their expertise in nanoengineering and material science. Also, I am lucky to have many supportive colleagues and friends in the lab. Anzu Kawazoe can always cheer me up with her lighthearted mood, unveiling the optimistic side of life to me. Zach Wells and Harald Schafer, beside research collaborations, brought me a lot of fun outside the lab, directing me to the ocean around Santa Barbara. I also very enjoyed the company of Yufei Hao, Taku Hachisu, Simone Fani, Jasper van der Lagemaat and Behnam Khojasteh during their visit to my lab. Starting from the time I joined the lab in Philadelphia, I have been gaining support from Marco Janko and Bin Li. I am really v grateful to have all of them as my colleagues. My friends outside the lab also helped me passing through the challenging paths toward the finish line of my PhD. I want to give special thanks to Mingwei Tang, my friend since childhood, who also obtained a PhD degree in the US. We traveled together to many sights and cities. Those trips reduced my anxiety and rebuilt my courage to march forward. I also value the friendship with Yan Kong and Zhenyu Yang, who provided me after-work entertainment in Harold Frank hall, and emotional support. I miss and appreciate the time with Pingge Jiang, Lei Song, Tao Sun, Yaou Zhou and Yanan Yang during my study in Philadelphia.I appreciate the love, care and support from my girlfriend, Huihui Zhang, during the last two years of my PhD. A delightful, sweet talk with her can always relieve my stress and hearten me, motivating me to look beyond the difficulties.Finally, I like to thank my family, especially my grandmother. I grew up bathing in her unconditional love. She was always proud of me and had the strongest belief in me, from preschool till PhD, granting me the faith to overcome any challenges. I deeply regret not being able to spend enough time with her at home.

show abstract

Speech enhancement using beamforming and non negative matrix factorization for robust speech recognition in the CHiME-3 challenge

Cited by 9 publications

References 8 publications

An analysis of environment, microphone and data simulation mismatches in robust speech recognition

An analysis of environment, microphone and data simulation mismatches in robust speech recognition

The third ‘CHiME’ speech separation and recognition challenge: Analysis and outcomes

Tactile Sensing, Information, and Feedback via Wave Propagation

Contact Info

Product

Resources

About