Robust speaker-independent word recognition using static, dynamic and acceleration features: experiments with Lombard and noisy speech

Hanson, Brian A.; Applebaum, Ted H.

doi:10.1109/icassp.1990.115973

Cited by 78 publications

(37 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Including temporal information with static features have shown to improve performance for automatic speech recognition (ASR) [29,30]. In this work we include temporal information via vector stacking.…”

Section: Temporal Informationmentioning

confidence: 99%

A Comparison of Perceptually Motivated Loss Functions for Binary Mask Estimation in Speech Separation

Websdale¹,

Milner²

2017

Interspeech 2017

View full text Add to dashboard Cite

This work proposes and compares perceptually motivated loss functions for deep learning based binary mask estimation for speech separation. Previous loss functions have focused on maximising classification accuracy of mask estimation but we now propose loss functions that aim to maximise the hit minus false-alarm (HIT-FA) rate which is known to correlate more closely to speech intelligibility. The baseline loss function is binary cross-entropy (CE), a standard loss function used in binary mask estimation, which maximises classification accuracy. We propose first a loss function that maximises the HIT-FA rate instead of classification accuracy. We then propose a second loss function that is a hybrid between CE and HIT-FA, providing a balance between classification accuracy and HIT-FA rate. Evaluations of the perceptually motivated loss functions with the GRID database show improvements to HIT-FA rate and ESTOI across babble and factory noises. Further tests then explore application of the perceptually motivated loss functions to a larger vocabulary dataset.

show abstract

Section: Temporal Informationmentioning

confidence: 99%

A Comparison of Perceptually Motivated Loss Functions for Binary Mask Estimation in Speech Separation

Websdale¹,

Milner²

2017

Interspeech 2017

View full text Add to dashboard Cite

show abstract

“…Yang et al [10] inspected that the functions of the basilar membrane in the human ear that can be viewed as an affine wavelet transform. Therefore, the filterbank used in auditory-based speech recognition systems [11][12][13] should also possess a wavelet property. However, in general, to construct the Continuous Wavelet Transform (CWT) that requires an infinite number of translations of the wavelet function; in this case, this proposed work only uses a finite set of filters for transformation.…”

Section: Introductionmentioning

confidence: 99%

A robust BFCC feature extraction for ASR system

Kuan

Tsai

Sung

et al. 2016

AIR

View full text Add to dashboard Cite

An auditory-based feature extraction algorithm naming the Basilar-membrane Frequency-band Cepstral Coefficient (BFCC) is proposed to increase the robustness for automatic speech recognition. Compared to Fourier spectrogram based of the MelFrequency Cepstral Coefficient (MFCC) method, the proposed BFCC method engages an auditory spectrogram based on a gammachirp wavelet transform to simulate the auditory response of human inner ear to improve the noise immunity. In addition, the Hidden Markov Model (HMM) is used for evaluating the proposed BFCC in phases of training and testing purposes conducted by AURORA-2 corpus with different Signal-to-Noise Ratios (SNRs) degrees of datasets. The experimental results indicate the proposed BFCC, compared with MFCC, Gammatone Wavelet Cepstral Coefficient (GWCC), and Gammatone Frequency Cepstral Coefficient (GFCC), improves the speech recognition rate by 13%, 17%, and 0.5% respectively, on average given speech samples with SNRs ranging from -5 to 20 dB.

show abstract

“…average spectral magnitude of speech in input signal reference spectral magnitude (4) where reference spectral magnitude is defined by a fixed value which is the average spectral magnitude of all words. Intensity variation factor G depends on spectral magnitude of the input word, and thus depends on the speaker, the type and level of noise.…”

Section: Restoration Of Clean Speech From Noisylombard Speechmentioning

confidence: 99%

“…The multi-style training method uses Lombard speech for training data [9]. The dynamic feature is known to be robust to Lombard speech recognition [4], and the codebook adaptation and acoustic phonetic variability models for HMM adaptation are used for Lombard effect compensation [10,11].…”

Section: Introductionmentioning

confidence: 99%

Lombard effect compensation and noise suppression for noisy Lombard speech recognition

Chi

Oh²

Proceeding of Fourth International Conference on Spoken Language Processing. ICSLP '96

View full text Add to dashboard Cite

The performance of speech recognition system degrades rapidly in the presence of ambient noise. To reduce the degradation, a degradation model is proposed which represents the spectral changes of speech signal uttered in noisy environments. The model uses frequency warping and amplitude scaling of each frequency band to simulate the variations of formant location, formant bandwidth, pitch, spectral tilt, and energy in each frequency band by Lombard effect. Another Lombard effect, the variation of overall vocal intensity is represented by a multiplicative constant term depending on spectral magnitude of input speech. The noise contamination is represented by an additive term in the frequency domain. According to this degradation model, the cepstral vector of clean speech is estimated from that of noisy-Lombard speech using spectral subtraction, spectral magnitude normalization, band-pass filter in LIN-LOG spectral domain, and multiple linear transformation. NoisyLombard speech data is collected by simulating the noisy environments using noises from automobile cabins, an exhibition hall, telephone booths in downtown, crowded streets, and computer rooms. The proposed method significantly reduces error rates in the recognition of 50 Korean word. For example, the recognition rate is 95.91% with this method, and 79.68% without this method at SNR (Signal-to-Noise Ratio) 10 dB.

show abstract

Robust speaker-independent word recognition using static, dynamic and acceleration features: experiments with Lombard and noisy speech

Cited by 78 publications

References 4 publications

A Comparison of Perceptually Motivated Loss Functions for Binary Mask Estimation in Speech Separation

A Comparison of Perceptually Motivated Loss Functions for Binary Mask Estimation in Speech Separation

A robust BFCC feature extraction for ASR system

Lombard effect compensation and noise suppression for noisy Lombard speech recognition

Contact Info

Product

Resources

About