• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.1
• /
• pp.71-76
• /
• 2023

This paper studied a method of attenuating nonlinear noise using a Wiener filter of a neural network structure in an acoustic noise attenuator. This system improves nonlinear noise attenuation performance with a deep learning algorithm using a neural network Wiener filter instead of using a conventional adaptive filter. A voice is estimated from a single input voice signal containing nonlinear noise using a 128-neuron, 8-neuron hidden layer and an error back propagation algorithm. In this study, a simulation program using the Keras library was written and a simulation was performed to verify the attenuation performance for nonlinear noise. As a result of the simulation, it can be seen that the noise attenuation performance of this system is significantly improved when the FNN filter is used instead of the Wiener filter even when nonlinear noise is included. This is because the complex structure of the FNN filter expresses any type of nonlinear characteristics well.

• The Journal of the Acoustical Society of Korea
• /
• v.13 no.2
• /
• pp.76-85
• /
• 1994

A speech enhancement algorithm which attenuates nonstationary colored noise is presented In this paper. The algorithm consists of a stationary Kalman filter and the simple speech/nonspeech detector. While the conventional enhancement systems are focused on a stationary and/or white background noise, this study Is focused on the mort realistic nonstationary and nonwhite noise. An AR model-based vector Kalman filter is used as a noise suppression system and a short-time energy threshold logic is used as a speech/nonspeech classifier. For Kalman filtering. noise coefficients are estimated in the nonspeech frame, and speech coefficients are estimated by applying the EM iteration algorithm. Simulation results using the car noise are presented based on the signal-to-noise ratio and informal listening tests. According to the experimental results, background noises in the nonspeech frames are eliminated almost completely, while some distortions are noticed in the speech frames. The distortion becomes severer as the SNR is reduced to 0dB and -5dB. Intelligibility, however, is not degraded significantly.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.6
• /
• pp.987-994
• /
• 2020

In this paper, we studied the effect of kernel size of CNN layer on performance in acoustic noise attenuators. This system uses a deep learning algorithm using a neural network adaptive prediction filter instead of using the existing adaptive filter. Speech is estimated from a single input speech signal containing noise using a 100-neuron, 16-filter CNN filter and an error back propagation algorithm. This is to use the quasi-periodic property in the voiced sound section of the voice signal. In this study, a simulation program using Tensorflow and Keras libraries was written and a simulation was performed to verify the performance of the noise attenuator for the kernel size. As a result of the simulation, when the kernel size is about 16, the MSE and MAE values are the smallest, and when the size is smaller or larger than 16, the MSE and MAE values increase. It can be seen that in the case of an speech signal, the features can be best captured when the kernel size is about 16.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.16 no.4
• /
• pp.625-632
• /
• 2021

This paper studies the effect of the number of filters in the CNN (Convolutional Neural Network) layer on the performance of a noise attenuator. Speech is estimated from a noised speech signal using a 64-neuron, 16-kernel CNN filter and an error back-propagation algorithm. In this study, in order to verify the performance of the noise attenuator with respect to the number of filters, a program using Keras library was written and simulation was performed. As a result of simulation, it can be seen that this system has the smallest MSE (Mean Squared Error) and MAE (Mean Absolute Error) values when the number of filters is 16, and the performance is the lowest when there are 4 filters. And when there are more than 8 filters, it was shown that the MSE and MAE values do not differ significantly depending on the number of filters. From these results, it can be seen that about 8 or more filters must be used to express the characteristics of the speech signal.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2009.05a
• /
• pp.160-163
• /
• 2009

최근 적외선 열 영상의 유용함이 발표되어감에 따라 이 기술에 대한 이해와 도입폭이 점진적으로 넓혀져 가고 있다. 국내는 물론 세계적으로 많은 제조, 개발 관련 업체들이 생겨나고 있으며, 업체들의 기술이 발전함에 따라 높은 온도 분해능과 해상도는 물론, 시스템 전체의 크기가 휴대가 가능할 정도로 소형화 되어가고 있는 추세이다. 이러한 열 영상 카메라에서 중요한 역할을 하는 적외선 열 감지 센서에서는 주변 온도, 대상 온도, 표면 온도 등의 측정 오차 및 열상 측정 소자의 온도, Bias 불안정 등의 매우 다양한 원인으로 인한 잡음이 발생하기 쉽다. 이러한 다양한 잡음의 감소는 해상도 및 온도 분해능의 향상과 직결되므로, 잡음을 줄이기 위한 많은 연구가 도처에서 행해지고 있다. 본 논문은 이러한 노력의 일환으로써 각각의 잡음 원인을 규명하지 않고 최종 열 영상 출력물에서 인접 프레임들의 비교, 혼합 하여 제거하는 3D Noise Reduction 기술을 이용해 노이즈를 감쇠하는 방법에 대해 연구하였다.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.6
• /
• pp.1103-1108
• /
• 2023

In this paper, we propose a new algorithm for attenuating the background noises in acoustic signal. This algorithm improves the noise attenuation performance by using the FNN(: Full-connected Neural Network) deep learning algorithm instead of the existing adaptive filter after wavelet transform. After wavelet transforming the input signal for each short-time period, noise is removed from a single input audio signal containing noise by using a 1024-1024-512-neuron FNN deep learning model. This transforms the time-domain voice signal into the time-frequency domain so that the noise characteristics are well expressed, and effectively predicts voice in a noisy environment through supervised learning using the conversion parameter of the pure voice signal for the conversion parameter. In order to verify the performance of the noise reduction system proposed in this study, a simulation program using Tensorflow and Keras libraries was written and a simulation was performed. As a result of the experiment, the proposed deep learning algorithm improved Mean Square Error (MSE) by 30% compared to the case of using the existing adaptive filter and by 20% compared to the case of using the STFT(: Short-Time Fourier Transform) transform effect was obtained.

• Proceedings of the Acoustical Society of Korea Conference
• /
• autumn
• /
• pp.163-166
• /
• 2001

초음파를 이용하여 인체 조직의 특성을 알아내는 방법은 매우 광범위하게 응용되어 오고있다. 그 중에서 초음파를 발생시킨 후 반사되어 되돌아오는 신호를 측정하여 그 감쇠 정도로부터 조직의 특성을 추정하는 방법이 많이 사용되고 있다. 이러한 감쇠현상은 발생된 초음파가 조직 내에서 흡수 또는 산란현상을 거치면서 주파수가 변이를 일으키기 때문에 발생한다. 따라서, 조직의 감쇠 특성을 알아내기 위해서, 주파수의 함수로 근사할 수 있는 감쇠 계수(attenuation coefficient)를 이용하여 시간에 따라 달라지는 주파수 변화를 추정한다. 그러나, 기존의 Ah(Auto-Regressive) 모델을 통한 시간영역 및 주파수 영역에서의 추정 방법을 사용하면 잡음이 존재하는 상황에서 시변 신호를 추정하는데 성능이 많이 저하된다. 본 논문에서는 이러한 단점을 보완하기 위해서, 가변 망각 인자와 재귀적인 TLS(Total Least Squares) 방법을 사용하여 시간에 따라 변하는 신호를 정확하게 추정하고 잡음환경에도 강인한 알고리듬을 제안하였다. 또한, 제안된 알고리듬은 추정 성능을 향상시킬 뿐 아니라 감쇠정도의 강약에 관계없이 망각인자의 값을 적응적으로 변화시켜 동작하는 장점을 가진다.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2007.11a
• /
• pp.777-777
• /
• 2007

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.08a
• /
• pp.432-437
• /
• 1998

NLMS 알고리즘을 채용한 음향반향제거기는 주변잡음에 대해서 적응필터의 계수가 오조정되어 반향제거기의 성능이 저하된다. 본 논문에서 음향반향제거기의 마이크 입력신호와 추정 오차신호의 상관도를 이용해서 주변 잡음신호에 의한 계수 오조정이 작은 적응 알고리즘과 잔여반향을 제거하기 위한 후처리기로 구성된 음향 반향 제거기를 제안한다. 기존의 NLMS 알고리즘이 입력신호의전력으로 적응상수를 정규화하지만 제안하는 알고리즘은 마이크 입력신호와 추정 오차신호의상관도와 입력신호 전력의 합으로 정규화한다. 적응필터가 반향 경로를 추정한 경우, 추정 오차신호에는 근단화자 신호가 대부분을 차지한다. 따라서 근단화자 신호가 있는 경우에는 상관도 값이 커져서 적응 상수가 작아지고 근단화자 신호에 의한 계수의 오조정을 줄일 수 있다. 후처리기도 마이크 입력신호와 추정 오차신호의 상관도를 마이크 입력신호의 전력으로 정규화한 값으로 추정 오차신호를 감쇠시킴으로써 근단화자 신호는 감쇠를 적게 하고 잔여반향을 감쇠시킨다. 멀티미디어 PC를 이용한 실험을 통해서 제안하는 알고리즘이 기존의 알고리즘에 비해서 우수한 성능을 보임을 확인했다.

• Journal of Advanced Navigation Technology
• /
• v.22 no.5
• /
• pp.351-358
• /
• 2018

In this study, we propose a method of suppressing the leakage noise signal of motor assembly through the test. The motor assembly is mounted on outside of the aircraft to rotate an antenna and must satisfy RE102 requirement on MIL-STD-461F in terms of electromagnetic interference. It is confirmed by RE102 test result hat the leakage noise signal of the equipment occurs due to external influx through the power and control cable and rotation of the motor. And it is ascertained that the part where internal/external physical shielding is difficult to rotate is the leakage path. To reduce the leakage noise signal, the electrical ground reinforcement and the electric shielding structure considering the operation of the equipment is applied and it is verified that the requirement is satisfied. Finally, we verified that required specification are met by applying circular corrugated choke with interlocking shapes and conductive grease to the noise leakage path.