Non-linear transformations of the feature space for robust speech recognition

Torre, De La; Segura, J.; Benítez, Domingo; Peinado,; Rubio,

doi:10.1109/icassp.2002.1005761

Cited by 14 publications

(12 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In (de la Torre et al, 2002;Skosan, 2005) it was shown that distortions caused by additive noise and linear filtering (distortions often encountered in telephone networks) corrupts spectral-based features (like those used in speech and speaker recognition systems) and, as a result, modifies their distributions-additive noise deforms the variances of these feature distributions while linear filtering alters the distribution means. In a previous contribution (Skosan and Mashao, 2004), we showed that a compensation technique, known as Histogram Equalization (HEQ), could be used to minimize the disparity between feature distributions collected in mismatched telephone environments.…”

Section: Introductionmentioning

confidence: 98%

Modified Segmental Histogram Equalization for robust speaker verification

Skosan

Mashao

2006

Pattern Recognition Letters

View full text Add to dashboard Cite

Section: Introductionmentioning

confidence: 98%

Modified Segmental Histogram Equalization for robust speaker verification

Skosan

Mashao

2006

Pattern Recognition Letters

View full text Add to dashboard Cite

“…Cepstral mean subtraction (CMS) is a simple but effective way to remove the dc component of features. Several other well-known normalization methods for feature domain have been proposed, such as cumulative histogram used in histogram equalization (HEQ) [6] and cepstral shape normalization (CSN) [7]. On the other hand, it is observed that human auditory system is remarkably more robust than the state-of-the-art ASR systems in the presence of variable noises.…”

Section: Introductionmentioning

confidence: 99%

Noise Robust Feature Scheme for Automatic Speech Recognition Based on Auditory Perceptual Mechanisms

Cai

Xiao

Pan

et al. 2012

IEICE Trans. Inf. & Syst.

View full text Add to dashboard Cite

SUMMARYMel Frequency Cepstral Coefficients (MFCC) are the most popular acoustic features used in automatic speech recognition (ASR), mainly because the coefficients capture the most useful information of the speech and fit well with the assumptions used in hidden Markov models. As is well known, MFCCs already employ several principles which have known counterparts in the peripheral properties of human hearing: decoupling across frequency, mel-warping of the frequency axis, logcompression of energy, etc. It is natural to introduce more mechanisms in the auditory periphery to improve the noise robustness of MFCC. In this paper, a k-nearest neighbors based frequency masking filter is proposed to reduce the audibility of spectra valleys which are sensitive to noise. Besides, Moore and Glasberg's critical band equivalent rectangular bandwidth (ERB) expression is utilized to determine the filter bandwidth. Furthermore, a new bandpass infinite impulse response (IIR) filter is proposed to imitate the temporal masking phenomenon of the human auditory system. These three auditory perceptual mechanisms are combined with the standard MFCC algorithm in order to investigate their effects on ASR performance, and a revised MFCC extraction scheme is presented. Recognition performances with the standard MFCC, RASTA perceptual linear prediction (RASTA-PLP) and the proposed feature extraction scheme are evaluated on a medium-vocabulary isolated-word recognition task and a more complex large vocabulary continuous speech recognition (LVCSR) task. Experimental results show that consistent robustness against background noise is achieved on these two tasks, and the proposed method outperforms both the standard MFCC and RASTA-PLP.

show abstract

“…1 The relative performance of MSE versus MTE scheme does not solely depend on the signal-to-noise ratio in the frequency band. 2 Higher-order derivatives of the input signal correspond to larger values of p, [21].…”

Section: Medium and Short-time Properties Of Energy Operatorsmentioning

confidence: 99%

“…The effect of noise on the features employed in a speech recognition front-end is nontrivial and can greatly influence the overall system performance. In this context, much work has been done minimizing this mismatch [2], [3] by using transformations of the noisy features to a "cleaner" feature domain, and thus improving their invariability to certain noise types. Other related work includes speech enhancement [4], normalization of the noisy features statistical properties [5]- [7], and dynamic feature combinations [8].…”

mentioning

confidence: 99%

On the Effects of Filterbank Design and Energy Computation on Robust Speech Recognition

Dimitriadis

Maragos

Potamianos

2011

IEEE Trans. Audio Speech Lang. Process.

View full text Add to dashboard Cite

Abstract-In this paper, we examine how energy computation and filterbank design contribute to the overall front-end robustness, especially when the investigated features are applied to noisy speech signals, in mismatched training-testing conditions. In prior work ("Auditory Teager energy cepstrum coefficients for robust speech recognition," D. Dimitriadis, P. Maragos, and A. Potamianos, in Proc. Eurospeech'05, Sep. 2005), a novel feature set called "Teager energy cepstrum coefficients" (TECCs) has been proposed, employing a dense, smooth filterbank and alternative energy computation schemes. TECCs were shown to be more robust to noise and exhibit improved performance compared to the widely used Mel frequency cepstral coefficients (MFCCs). In this paper, we attempt to interpret these results using a combined theoretical and experimental analysis framework. Specifically, we investigate in detail the connection between the filterbank design, i.e., the filter shape and bandwidth, the energy estimation scheme and the automatic speech recognition (ASR) performance under a variety of additive and/or convolutional noise conditions. For this purpose: 1) the performance of filterbanks using triangular, Gabor, and Gammatone filters with various bandwidths and filter positions are examined under different noisy speech recognition tasks, and 2) the squared amplitude and Teager-Kaiser energy operators are compared as two alternative approaches of computing the signal energy. Our end-goal is to understand how to select the most efficient filterbank and energy computation scheme that are maximally robust under both clean and noisy recording conditions. Theoretical and experimental results show that: 1) the filter bandwidth is one of the most important factors affecting speech recognition performance in noise, while the shape of the filter is of secondary importance, and 2) the Teager-Kaiser operator outperforms (on the average and for most noise types) the squared amplitude energy computation scheme for speech recognition in noisy conditions, especially, for large filter bandwidths. Experimental results show that selecting the appropriate filterbank and energy computation scheme can lead to significant error rate reduction over both MFCC and perceptual linear predicion (PLP) features for a variety of speech recognition tasks. A relative error rate reduction of up to 30% for MFCCs and 39% for PLPs is shown for the Aurora-3 Spanish Task.

show abstract

Non-linear transformations of the feature space for robust speech recognition

Cited by 14 publications

References 0 publications

Modified Segmental Histogram Equalization for robust speaker verification

Modified Segmental Histogram Equalization for robust speaker verification

Noise Robust Feature Scheme for Automatic Speech Recognition Based on Auditory Perceptual Mechanisms

On the Effects of Filterbank Design and Energy Computation on Robust Speech Recognition

Contact Info

Product

Resources

About