A minimum-mean-square-error noise reduction algorithm on Mel-frequency cepstra for robust speech recognition

By Dong Yu, Li Deng, Jasha Droppo, Jian Wu, Yifan Gong, A. Acero. 2008

92 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Dong Yu
2. Li Deng
3. Jinyu Li
4. Yifan Gong
5. Shinji Watanabe
6. M. Seltzer
7. Deliang Wang
8. Yoshua Bengio

In Signal Processing

1. Dong Yu
2. Jinyu Li
3. Li Deng
4. Yifan Gong
5. Deliang Wang
6. M. Seltzer
7. Shinji Watanabe
8. Jasha Droppo

Citations in the corpus

• SEGAN: Speech Enhancement Generative Adversarial Network (2017, methodology): “Speech enhancement has also been successfully applied as a preprocessing stage in speech recognition and speaker identifi-cation systems [3, 4, 5].”
• The design and collection of COSINE, a multi-microphone in situ speech corpus recorded in noisy environments (2012, background): “For example, mean/variance normalization, feature smoothing [3], and a variety of other feature cleaning/enhancement techniques show improvement over standard MFCC/PLP features [4, 5, 6].”
• Recurrent deep neural networks for robust speech recognition (2014, methodology): “To deal with the mismatched acoustical conditions between training and testing, feature space compensation approaches typically involve removing additive noise and channel distortions using speech enhancement techniques [1] such as spectral subtraction, Weiner filtering and MMSE estimators [2, 3, 4]”
• Very deep convolutional neural networks for robust speech recognition (2016, background): “Experimental results on the noisy Aurora4 task show that very promising performance can be achieved with the proposed model, even without using front-end de-noising [16] or sequence training [17].”
• An investigation of deep neural networks for noise robust speech recognition (2013, methodology): “In this work, we use an feature enhancement algorithm based on the Cepstral-domain Minimum Mean Squared Error (C-MMSE) criterion [2].”
• Improved cepstra minimum-mean-square-error noise reduction algorithm for robust speech recognition (2017, methodology): “Particularly, we work on improving a robust front-end named cepstra minimum mean square error (CMMSE) [22] which is very effective in dealing with noise when using the GMM-based acoustic models.”
• Robust Speech Recognition Using a Cepstral Minimum-Mean-Square-Error-Motivated Noise Suppressor (2008, background): “For subjective human listening, noise reduction has been Manuscript received October 17, 2007; revised February 22, 2008.”