• The Journal of the Acoustical Society of Korea
• /
• v.16 no.3E
• /
• pp.14-18
• /
• 1997

One fo the major problems in speech recognition is the mismatch between training and testing environments. Recently, SNR normalization technique, which normalizes the dynamic range of frequency channels in mel-scaled filterbank, was proposed[1]. While it showed improved robustness against additive noise, it requires a reliable speech detection mechanism and several adaptation parameters to be optimized. In this paper, we propose a modified SNR normalization technique. In this technique, we take simply the maximum of filterbank output and predetermined masking constant for each frequency band. According to the speaker-independent isolated word recognition in car noise environments, proposed modification yields better recognition performance that the original SNR normalization method, with rather reduced complexity.

• The Journal of the Acoustical Society of Korea
• /
• v.15 no.5
• /
• pp.90-98
• /
• 1996

An auditory model is proposed for robust speech recognition in noisy environments. The model consists of cochlear bandpass filters and nonlinear stages, and represents frequency and intensity information efficiently even in noisy environments. Frequency information of the signal is obtained by zero-crossing intervals, and intensity information is also incorporated by peak detectors and saturating nonlinearities. Also, the robustness of the zero-crossings in estimating frequency is verified by the developed analytic relationship of the variance of the level-crossing interval perturbations as a function of the crossing level values. The proposed auditory model is computationally efficient and free from many unknown parameters compared with other auditory models. Speaker-independent speech recognition experiments demonstrate the robustness of the proposed method.

• Journal of information and communication convergence engineering
• /
• v.1 no.4
• /
• pp.205-208
• /
• 2003

Voice activity detection is a important problem in the speech recognition and speech communication. This paper introduces new feature parameter which are reconstructed by spectral entropy of information theory for robust voice activity detection in the noise environment, then analyzes and compares it with energy method of voice activity detection and performance. In experiments, we confirmed that spectral entropy and its reconstructed parameter are superior than the energy method for robust voice activity detection in the various noise environment.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.41 no.3
• /
• pp.67-77
• /
• 2004

The recent research has been focused on fusion of audio and visual features for reliable speech recognition in noisy environments. In this paper, we propose a neural network based model of robust speech recognition by integrating audio, visual, and contextual information. Bimodal Neural Network(BMNN) is a multi-layer perception of 4 layers, each of which performs a certain level of abstraction of input features. In BMNN the third layer combines audio md visual features of speech to compensate loss of audio information caused by noise. In order to improve the accuracy of speech recognition in noisy environments, we also propose a post-processing based on contextual information which are sequential patterns of words spoken by a user. Our experimental results show that our model outperforms any single mode models. Particularly, when we use the contextual information, we can obtain over 90% recognition accuracy even in noisy environments, which is a significant improvement compared with the state of art in speech recognition. Our research demonstrates that diverse sources of information need to be integrated to improve the accuracy of speech recognition particularly in noisy environments.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.7
• /
• pp.603-611
• /
• 2003

This paper proposes an effective feature compensation scheme based on speech model for achieving robust speech recognition. Conventional model-based method requires off-line training with noisy speech database and is not suitable for online adaptation. In the proposed scheme, we can relax the off-line training with noisy speech database by employing the parallel model combination technique for estimation of correction factors. Applying the model combination process over to the mixture model alone as opposed to entire HMM makes the online model combination possible. Exploiting the availability of noise model from off-line sources, we accomplish the online adaptation via MAP (Maximum A Posteriori) estimation. In addition, the online channel estimation procedure is induced within the proposed framework. For more efficient implementation, we propose a selective model combination which leads to reduction or the computational complexities. The representative experimental results indicate that the suggested algorithm is effective in realizing robust speech recognition under the combined adverse conditions of additive background noise and channel distortion.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.2
• /
• pp.160-166
• /
• 2002

In this paper, we implement a new vocabulary-independent speech recognition system that uses CV, VCCV, VC recognition unit. Since these recognition units are extracted in the trowel region of syllable, the segmentation is easy and robust. And in the case of not existing VCCV unit, the units are replaced by combining VC and CV semi-syllable model. Clustering of vowel group and applying combination rule to the substitution model in the case of not existing of VCCV model lead to 5.2% recognition performance improvement from 90.4% (Model A) to 95.6% (Model C) in the first candidate. The recognition results that is 98.8% recognition rate in the second candidate confirm the effectiveness of the proposed method.

• The Journal of the Acoustical Society of Korea
• /
• v.41 no.3
• /
• pp.326-334
• /
• 2022

In this paper, we propose speech enhancement algorithm as a pre-processing for robust speech recognition in noisy environments. Auxiliary-function-based Independent Vector Analysis (AuxIVA) is performed with weighted covariance matrix using time-varying variances with scaling factor from target masks representing time-frequency contributions of target speech. The mask estimates can be obtained using Neural Network (NN) pre-trained for speech extraction or diffuseness using Coherence-to-Diffuse power Ratio (CDR) to find the direct sounds component of a target speech. In addition, outputs for omni-directional noise are closely chained by sharing the time-varying variances similarly to independent subspace analysis or IVA. The speech extraction method based on AuxIVA is also performed in Independent Low-Rank Matrix Analysis (ILRMA) framework by extending the Non-negative Matrix Factorization (NMF) for noise outputs to Non-negative Tensor Factorization (NTF) to maintain the inter-channel dependency in noise output channels. Experimental results on the CHiME-4 datasets demonstrate the effectiveness of the presented algorithms.

• Proceedings of the KSPS conference
• /
• 2003.10a
• /
• pp.149-152
• /
• 2003

In this paper, data-driven temporal filter methods[1] are investigated for robust feature extraction. A principal component analysis technique is applied to the time trajectories of feature sequences of training speech data to get appropriate temporal filters. We did recognition experiments on the Korean connected digit telephone speech database released by SITEC, with data-driven temporal filters. Experimental results are discussed with our findings.

• Speech Sciences
• /
• v.2
• /
• pp.57-66
• /
• 1997

This paper studies the effects of the method, CMS(Cepstral Mean Subtraction), (which compensates for some of the speech distortion. caused by telephone channels), on the performance of the text-dependent speaker identification system. This system is based on the VQ(Vector Quantization) and HMM(Hidden Markov Model) method and chooses the LPC-Cepstrum and Mel-Cepstrum as the feature vectors extracted from the speech data transmitted through telephone channels. Accordingly, we can compare the correct recognition rates of the speaker identification system between the use of LPC-Cepstrum and Mel-Cepstrum. Finally, from the experiment results table, it is found that the Mel-Cepstrum parameter is proven to be superior to the LPC-Cepstrum and that recognition performance improves by about 10% when compensating for telephone channel using the CMS.

• Journal of Korea Society of Industrial Information Systems
• /
• v.6 no.4
• /
• pp.81-88
• /
• 2001

In this paper, a Korean digit recognition system based on a multilayer Perceptron is implemented. We also investigate the performance of widely used speech features, such as the Mel-scale filterbank, MFCC, LPCC, and PLP coefficients, by applying them as input of the proposed recognition system. In order to build a robust speech system, the experiments for demonstrating its recognition performance for the clean data as well as corrupt data are carried out. In experiments of recognizing 20 Korean digit, we found that the Mel-scale filterbank coefficients performs best in terms of recognition accuracy for the speech dependent and speech independent database even though noise is considerably added.