• The Journal of the Acoustical Society of Korea
• /
• v.22 no.2E
• /
• pp.62-68
• /
• 2003

This paper addresses a novel noise-compensation scheme to solve the mismatch problem between training and testing condition for the automatic speech recognition (ASR) system, specifically in car environment. The conventional spectral subtraction schemes rely on the signal-to-noise ratio (SNR) such that attenuation is imposed on that part of the spectrum that appears to have low SNR, and accentuation is made on that part of high SNR. However, these schemes are based on the postulation that the power spectrum of noise is in general at the lower level in magnitude than that of speech. Therefore, while such postulation is adequate for high SNR environment, it is grossly inadequate for low SNR scenarios such as that of car environment. This paper proposes an efficient spectral subtraction scheme focused specifically to low SNR noisy environment by extracting harmonics distinctively in speech spectrum. Representative experiments confirm the superior performance of the proposed method over conventional methods. The experiments are conducted using car noise-corrupted utterances of Aurora2 corpus.

• Proceedings of the IEEK Conference
• /
• 2006.06a
• /
• pp.357-358
• /
• 2006

We propose the robust PBFLMS in noises, which is the enhanced acoustic echo canceller using ACPBF-LMS(Alternative Constrained Partitioned Block Frequency domain Least Mean Square) algorithm. The defect of the block structure filtering is the deterioration of convergence efficiency from noise and interference. To improve the performance of convergence efficiency, noise effect should be reduced. The new method of reducing noise effect is proposed, which apply the estimated background noise to adaptive filter step size. By experiments, the proposed acoustic echo canceller has TCL of 50dB, and always provides faster convergence speed and lower complexity than the full-tap NLMS. We also carried out an implementation of PBFLMS using ARM9EJ-S.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2009.10a
• /
• pp.368-369
• /
• 2009

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2012.04a
• /
• pp.212-213
• /
• 2012

• The Journal of the Acoustical Society of Korea
• /
• v.18 no.7
• /
• pp.112-117
• /
• 1999

In case of spectral subtraction for noise robust speech recognition system, this method often makes loss of speech signal. In this study, we propose a method that variation and determination of Gaussian function at semi-continuous HMM(Hidden Markov Model) is made on the basis of SNR criteria function, in which SNR means signal to noise ratio between estimation noise and subtracted signal per frame. For proving effectiveness of this method, we show the estimation error to be related with the magnitude of estimated noise through signal waveform. For this reason, Gaussian function is varied and determined by SNR. When we test recognition rate by computer simulation under the noise environment of driving car over the speed of 80㎞/h, the proposed Gaussian decision method by SNR turns out to get more improved recognition rate compared with the frequency subtracted and non-subtracted cases.

• The Journal of the Acoustical Society of Korea
• /
• v.43 no.4
• /
• pp.406-413
• /
• 2024

The proposed method uses Generalized Cross Correlation - Phase Transform (GCC-PHAT) method with nonlinear preprocessing and a frequency domain low-pass filter. In this paper, by reinterpreting the calculation process of GCC-PHAT as DFT, we derive that there is an effective frequency band used for time delay estimation in GCC-PHAT, and by using only the effective band using a low-pass filter, the noise component is reduced and it improvesthe time delay performance in impulsive noise environments. By comparing the proposed method with the traditional GCC-PHAT in an impulsive noise environment, we show that the GCC-PHAT becomes more robust to the impulsive noise.

• ETRI Journal
• /
• v.35 no.3
• /
• pp.512-522
• /
• 2013

In this paper, the radius weight mean (RWM) and the feature-embedding technique are used to present a novel watermarking scheme for color images. Simulations validate that the stego-images generated by the proposed scheme are robust against most common image-processing operations, such as compression, color quantization, bit truncation, noise addition, cropping, blurring, mosaicking, zigzagging, inversion, (edge) sharpening, and so on. The proposed method possesses outstanding performance in resisting high compression ratio attacks: JPEG2000 and JPEG. Further, to provide extra hiding storage, a steganographic method using the RWM with the least significant bit substitution technique is suggested. Experiment results indicate that the resulting perceived quality is desirable, whereas the peak signal-to-noise ratio is high. The payload generated using the proposed method is also superior to that generated by existing approaches.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2005.05a
• /
• pp.840-845
• /
• 2005

This paper presents a robust optimal design method for a periodic structure type of MEMS resonator that is vulnerable to mode localization. The robust configuration of such a MEMS resonator to fabrication error is implemented by changing the regularity of periodic structure. For the mathematical convenience, the MEMS resonator is first modeled as a multi pendulum system. The index representing the measure of mode variation is then introduced using the perturbation method and the concept of modal assurance criterion. Finally, the optimal intentional mistuning, minimizing the expectation of the irregularity measure for each substructure, is determined for the normal distributed fabrication error and its robustness in the design of MEMS resonator to the fabrication error is demonstrated with numerical examples.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.15 no.8 s.101
• /
• pp.931-938
• /
• 2005

This paper presents a robust optimal design method for a periodic structure type of MEMS resonator that is vulnerable to mode localization. The robust configuration of such a MEMS resonator to fabrication error is implemented by changing the regularity of periodic structure For the mathematical convenience, the MEMS resonator is first modeled as a multi-pendulum system. The index representing the measure of mode variation is then introduced using the perturbation method and the concept of modal assurance criterion. Finally, the optimal intentional mistuning, minimizing the expectation of the irregularity measure for each substructure, is determined for the normal distributed fabrication error and its robustness in the design of MEMS resonator to the fabrication error is demonstrated with numerical examples.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2006.11a
• /
• pp.925-933
• /
• 2006

Using the physics-based model for the vibration isolation system, the model uncertainties are described. With the model including parameter perturbations, the robust controller to meet the robust performance and stability is designed through $\mu$-synthesis by DK-iteration. The order of controller is reduced by virtue of Hankel norm approximation technique to allow the efficient implementation in the real-time experimental environment without any performance degradation. The performance of the reduced $\mu$-controller is accessed in comparison with the original one. The experiments validate the superiority of the proposed control scheme against the model uncertainties and its applicability with varying payload.