A Perceptually Inspired Data Augmentation Method for Noise Robust CNN Acoustic Models

Tóth, László; Kovács, George L.; Compernolle, Dirk Van

doi:10.1007/978-3-319-99579-3_71

Cited by 18 publications

(13 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On the other hand, both the size and position of the frequency masks in SpecAugment are chosen stochastically, and differ for every input in the minibatch. More ideas for structurally omitting frequency data of spectrograms have been discussed in [50].…”

Section: Discussionmentioning

confidence: 99%

SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition

et al. 2019

View full text Add to dashboard Cite

We present SpecAugment, a simple data augmentation method for speech recognition. SpecAugment is applied directly to the feature inputs of a neural network (i.e., filter bank coefficients). The augmentation policy consists of warping the features, masking blocks of frequency channels, and masking blocks of time steps. We apply SpecAugment on Listen, Attend and Spell networks for end-to-end speech recognition tasks. We achieve state-of-the-art performance on the LibriSpeech 960h and Swichboard 300h tasks, outperforming all prior work. On LibriSpeech, we achieve 6.8% WER on test-other without the use of a language model, and 5.8% WER with shallow fusion with a language model. This compares to the previous stateof-the-art hybrid system of 7.5% WER. For Switchboard, we achieve 7.2%/14.6% on the Switchboard/CallHome portion of the Hub5'00 test set without the use of a language model, and 6.8%/14.1% with shallow fusion, which compares to the previous state-of-the-art hybrid system at 8.3%/17.3% WER.

show abstract

Section: Discussionmentioning

confidence: 99%

SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition

et al. 2019

View full text Add to dashboard Cite

show abstract

“…DNN with 26.8% examples-based enhancement [28] CNN with channel dropout [20] 26.8% CNN with data augmentation [29] 25.6% GMM with 25.5% auditory spectral enhancement [30] TDNN with Gabor filters, 25.0% multi-band processing and channel dropout [19] Current paper 24.3%…”

Section: Methods Wersmentioning

confidence: 99%

Examining the Combination of Multi-Band Processing and Channel Dropout for Robust Speech Recognition

Kovács¹,

Tóth²,

Compernolle³

et al. 2019

Interspeech 2019

Self Cite

View full text Add to dashboard Cite

A pivotal question in Automatic Speech Recognition (ASR) is the robustness of the trained models. In this study, we investigate the combination of two methods commonly applied to increase the robustness of ASR systems. On the one hand, inspired by auditory experiments and signal processing considerations, multi-band band processing has been used for decades to improve the noise robustness of speech recognition. On the other hand, dropout is a commonly used regularization technique to prevent overfitting by keeping the model from becoming over-reliant on a small set of neurons. We hypothesize that the careful combination of the two approaches would lead to increased robustness, by preventing the resulting model from over-rely on any given band. To verify our hypothesis, we investigate various approaches for the combination of the two methods using the Aurora-4 corpus. The results obtained corroborate our initial assumption, and show that the proper combination of the two techniques leads to increased robustness, and to significantly lower word error rates (WERs). Furthermore, we find that the accuracy scores attained here compare favourably to those reported recently on the clean training scenario of the Aurora-4 corpus.

show abstract

“…However, as a side effect, deep learning-based models require a large amount of labeled training data to combat overfitting and ensure high accuracy, especially for speech recognition tasks with few training data. Therefore, a lot of data augmentation methods for ASR [7,8,9,10,11,12] were proposed, mainly on augmenting speech data. For example, speed perturbation [7], pitch adjust [8], adding noise [9] and vocal tract length perturbation increases the quantity of speech data by adjusting the speed or pitch of the audio, or by adding noisy audio on the original clean audio, or by transforming spectrograms.…”

Section: Introductionmentioning

confidence: 99%

“…Therefore, a lot of data augmentation methods for ASR [7,8,9,10,11,12] were proposed, mainly on augmenting speech data. For example, speed perturbation [7], pitch adjust [8], adding noise [9] and vocal tract length perturbation increases the quantity of speech data by adjusting the speed or pitch of the audio, or by adding noisy audio on the original clean audio, or by transforming spectrograms. Recently, SpecAugment [10] was proposed to mask the mel-spectrogram along the time and frequency axes, and achieve good improvements on recognition accuracy.…”

Section: Introductionmentioning

confidence: 99%

MixSpeech: Data Augmentation for Low-Resource Automatic Speech Recognition

Meng

Tan

et al. 2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

In this paper, we propose MixSpeech, a simple yet effective data augmentation method based on mixup for automatic speech recognition (ASR). MixSpeech trains an ASR model by taking a weighted combination of two different speech features (e.g., mel-spectrograms or MFCC) as the input, and recognizing both text sequences, where the two recognition losses use the same combination weight. We apply MixSpeech on two popular end-to-end speech recognition models including LAS (Listen, Attend and Spell) and Transformer, and conduct experiments on several low-resource datasets including TIMIT, WSJ, and HKUST. Experimental results show that MixSpeech achieves better accuracy than the baseline models without data augmentation, and outperforms a strong data augmentation method SpecAugment on these recognition tasks. Specifically, MixSpeech outperforms SpecAugment with a relative PER improvement of 10.6% on TIMIT dataset, and achieves a strong WER of 4.7% on WSJ dataset.

show abstract

A Perceptually Inspired Data Augmentation Method for Noise Robust CNN Acoustic Models

Cited by 18 publications

References 26 publications

SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition

SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition

Examining the Combination of Multi-Band Processing and Channel Dropout for Robust Speech Recognition

MixSpeech: Data Augmentation for Low-Resource Automatic Speech Recognition

Contact Info

Product

Resources

About