Ideal ratio mask estimation using deep neural networks for robust speech recognition

Narayanan, Arun; Wang, DeLiang

doi:10.1109/icassp.2013.6639038

Cited by 481 publications

(270 citation statements)

References 19 publications

Supporting

Mentioning

263

Contrasting

Unclassified

Order By: Relevance

“…Besides getting the mapping feature directly, the DNN can also be used to train an ideal binary mask (IBM) which can be used to separate the clean speech from background noise as shown in Figure 10 [91,177,178]. With a priori knowledge of noise types and SNR, we can generate IBMs as training targets and use noisy power spectral as training data.…”

Section: Speech Recognition and Verification For The Internet Ofmentioning

confidence: 99%

A Survey on Machine Learning‐Based Mobile Big Data Analysis: Challenges and Applications

Xie

Song

et al. 2018

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

This paper attempts to identify the requirement and the development of machine learning-based mobile big data (MBD) analysis through discussing the insights of challenges in the mobile big data. Furthermore, it reviews the state-of-the-art applications of data analysis in the area of MBD. Firstly, we introduce the development of MBD. Secondly, the frequently applied data analysis methods are reviewed. Three typical applications of MBD analysis, namely, wireless channel modeling, human online and offline behavior analysis, and speech recognition in the Internet of Vehicles, are introduced, respectively. Finally, we summarize the main challenges and future development directions of mobile big data analysis.

show abstract

Section: Speech Recognition and Verification For The Internet Ofmentioning

confidence: 99%

A Survey on Machine Learning‐Based Mobile Big Data Analysis: Challenges and Applications

Xie

Song

et al. 2018

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

show abstract

“…For robust ASR, the ideal ratio mask (IRM), defined as the ratio of speech energy to total energy (speech and noise) in each time-frequency unit, has been shown to have a better performance compared to using IBM in a large vocabulary speech recognition task [39]. In [40], the DNN is used to estimate the instantaneous SNR for computing IRM, subsequently applied to filter out noise from a noisy Mel spectrogram. The recurrent neural networks (RNNs), with their ability to model the temporal dependencies in speech, have also been employed to estimate the time-frequency masks from the magnitude spectrum of a noisy signal for speech enhancement and recognition [41].…”

Section: Speech Enhancement Using Dnnmentioning

confidence: 99%

Feature mapping using far-field microphones for distant speech recognition

Himawan

Motlíček

Sridharan

2016

Speech Communication

View full text Add to dashboard Cite

Acoustic modeling based on deep architectures has recently gained remarkable success, with substantial improvement of speech recognition accuracy in several automatic speech recognition (ASR) tasks. For distant speech recognition, the multi-channel deep neural network based approaches rely on the powerful modeling capability of deep neural network (DNN) to learn suitable representation of distant speech directly from its multi-channel source. In this model-based combination of multiple microphones, features from each channel are concatenated and used together as an input to DNN. This allows integrating the multi-channel audio for acoustic modeling without any pre-processing steps. Despite powerful modeling capabilities of DNN, an environmental mismatch due to noise and reverberation may result in severe performance degradation when features are simply fed to a DNN without a feature enhancement step. In this paper, we introduce the nonlinear bottleneck feature mapping approach using DNN, to transform the noisy and reverberant features to its clean version. The bottleneck features trained on clean signal are used as a teacher signal because they contain relevant information to phoneme classification, and the mapping is performed with the objective of suppressing noise and reverberation. The individual and combined impacts of beamforming and speaker adaptation techniques along with the feature mapping are examined for distant large vocabulary speech recognition, using a single and multiple far-field microphones. As an alternative to beamforming, experiments with concatenating multiple channel features are conducted. The experimental results on the AMI meeting corpus show that the feature mapping, used in combination with beamforming and speaker adaptation yields a distant speech recognition performance below 50% word error rate (WER), using DNN for acoustic modeling.

show abstract

“…The IRM is widely used in speech segregation, speech enhancement and noise robust ASR [13], [15], [17], [18]. The IRM is defined as follows: where S 2 (t, f ) and N 2 (t, f ) denote the speech and noise energy at a particular T-F point, respectively.…”

Section: A Irm Featuresmentioning

confidence: 99%

Effectiveness of ideal ratio mask for non-intrusive quality assessment of noise suppressed speech

Soni

Patil

2017

2017 25th European Signal Processing Conference (EUSIPCO)

View full text Add to dashboard Cite

Abstract-The Ideal Ratio Mask (IRM) has proven to be very effective tool in many applications such as speech segregation, speech enhancement for hearing aid design and noise robust speech recognition tasks. The IRM provides information regarding the amount of signal power at each Time-Frequency (T-F) unit in a given signal-plus-noise mixture. In this paper, we propose to use the IRM for non-intrusive quality assessment of noise suppressed speech. Since the quality of noise suppressed speech is dependent on the residual noise present in speech, IRM can be extremely useful for its quality assessment. The quality assessment problem is posed as a regression problem and the mapping between statistics of acoustic features, namely, Mel Filterbank Energies (FBEs) plus IRM features and the subjective score of the corresponding utterances was found using single-layer Artificial Neural Network (ANN). The results of our experiments suggest that by using the mean of FBEs and IRM features as the input, the quality prediction accuracy was significantly increased.

show abstract

Ideal ratio mask estimation using deep neural networks for robust speech recognition

Cited by 481 publications

References 19 publications

A Survey on Machine Learning‐Based Mobile Big Data Analysis: Challenges and Applications

A Survey on Machine Learning‐Based Mobile Big Data Analysis: Challenges and Applications

Feature mapping using far-field microphones for distant speech recognition

Effectiveness of ideal ratio mask for non-intrusive quality assessment of noise suppressed speech

Contact Info

Product

Resources

About