Smoothed Nonlinear Energy Operator-Based Amplitude Modulation Features for Robust Speech Recognition

Alam, Md. Jahangir; Kenny, Patrick; O’Shaughnessy, D.

doi:10.1007/978-3-642-38847-7_22

Cited by 5 publications

(1 citation statement)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this work, we present the Smoothed Nonlinear Energy Operator (SNEO) [4] as a shock outcome predictor. SNEO is based on the analysis of the local energy content of the VF-waveform and in this work we also compare it to other classical shock outcome prediction features.…”

Section: Introductionmentioning

confidence: 99%

Nonlinear Energy Operators for Defibrillation Shock Outcome Prediction

Chicote

Irusta

Aramendi

et al. 2016

2016 Computing in Cardiology Conference (CinC)

View full text Add to dashboard Cite

Accurate prediction of shock success would avoid futile defibrillation attempts that may damage the myocardium, and would help optimizing treatment decisions for out-of-hospital cardiac arrest (OHCA) patients. This work applies the Smoothed Nonlinear Energy Operator (SNEO) to analyze the energy content of the pre-shock ventricular fibrillation (VF) waveform adquired by automated external defibrillators (AED).A database of 419 shocks was analyzed and shock outcome predictors were calculated in the a 5-s pre-shock ECG segment. The SNEO was compared to some classical VF features. For each feature a detector of successful shocks was designed minimizing the Balanced Error Rate (BER). Finally, using SNEO as shock outcome predictor the minimun pre-shock segment duration was determined. The SNEO has proven to be a good shock outcome predictor even for 2-s segments and it could be used to optimize treatment decisions for OHCA patients.

show abstract

Section: Introductionmentioning

confidence: 99%

Nonlinear Energy Operators for Defibrillation Shock Outcome Prediction

Chicote

Irusta

Aramendi

et al. 2016

2016 Computing in Cardiology Conference (CinC)

View full text Add to dashboard Cite

show abstract

Robust feature extraction based on an asymmetric level-dependent auditory filterbank and a subband spectrum enhancement technique

Alam

Kenny

O’Shaughnessy

2014

Digital Signal Processing

View full text Add to dashboard Cite

Regularized minimum variance distortionless response-based cepstral features for robust continuous speech recognition

Alam

Kenny

O’Shaughnessy

2015

Speech Communication

Self Cite

View full text Add to dashboard Cite

12In this paper, we present robust feature extractors that incorporate a regularized minimum variance distortionless response 13 (RMVDR) spectrum estimator instead of the discrete Fourier transform-based direct spectrum estimator, used in many front-ends 14 including the conventional MFCC, to estimate the speech power spectrum. Direct spectrum estimators, e.g., single tapered periodogram, 15 have high variance and they perform poorly under noisy and adverse conditions. To reduce this performance drop we propose to increase 16 the robustness of speech recognition systems by extracting features that are more robust based on the regularized MVDR technique. The 17 RMVDR spectrum estimator has low spectral variance and is robust to mismatch conditions. Based on the RMVDR spectrum estima-18 tor, robust acoustic front-ends, namely, are regularized MVDR-based cepstral coefficients (RMCC), robust RMVDR cepstral coeffi-19 cients (RRMCC) and normalized RMVDR cepstral coefficients (NRMCC). In addition to the RMVDR spectrum estimator, 20 RRMCC and NRMCC also utilize auditory domain spectrum enhancement methods, auditory spectrum enhancement (ASE) and 21 medium duration power bias subtraction (MDPBS) techniques, respectively, to improve the robustness of the feature extraction method. 22 Experimental speech recognition results are conducted on the AURORA-4 large vocabulary continuous speech recognition corpus and 23 performances are compared with the Mel frequency cepstral coefficients (MFCC), perceptual linear prediction (PLP), MVDR spectrum 24 estimator-based MFCC, perceptual MVDR (PMVDR), cochlear filterbank cepstral coefficients (CFCC), power normalized cepstral 25 coefficients (PNCC), ETSI advancement front-end (ETSI-AFE), and the robust feature extractor (RFE) of Alam et al. (2012). 26Experimental results demonstrate that the proposed robust feature extractors outperformed the other robust front-ends in terms of 27 percentage word error rate on the AURORA-4 large vocabulary continuous speech recognition (LVCSR) task under clean and 28 multi-condition training conditions. In clean training conditions, on average, the RRMCC and NRMCC provide significant reductions 29 in word error rate over the rest of the front-ends. In multi-condition training, the RMCC, RRMCC, and NRMCC perform slightly 30 better in terms of the average word error rate than the rest of the front-ends used in this work. 31 Ó 2015 Published by Elsevier B.V.32 33

show abstract

Smoothed Nonlinear Energy Operator-Based Amplitude Modulation Features for Robust Speech Recognition

Cited by 5 publications

References 21 publications

Nonlinear Energy Operators for Defibrillation Shock Outcome Prediction

Nonlinear Energy Operators for Defibrillation Shock Outcome Prediction

Robust feature extraction based on an asymmetric level-dependent auditory filterbank and a subband spectrum enhancement technique

Regularized minimum variance distortionless response-based cepstral features for robust continuous speech recognition

Contact Info

Product

Resources

About