The Journal of the Korea institute of electronic communication sciences (한국전자통신학회논문지)

Korea Institute of Electronic Communication Science (한국전자통신학회)
권호 표지
DOI QR코드

DOI QR Code

Noise Canceler Based on Deep Learning Using Discrete Wavelet Transform

이산 Wavelet 변환을 이용한 딥러닝 기반 잡음제거기

  • Haeng-Woo Lee (Dept. Intelligent Information Communication Engineering, Namseoul University)
  • 이행우 (남서울대학교 지능정보통신공학과)
  • Received : 2023.10.11
  • Accepted : 2023.12.27
  • Published : 2023.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

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.

본 논문에서는 음향신호의 배경잡음을 감쇠하기 위한 새로운 알고리즘을 제안한다. 이 알고리즘은 이산 웨이블릿 변환(DWT: Discrete Wavelet Transform) 후 기존의 적응필터를 대신 FNN(: Full-connected Neural Network) 심층학습 알고리즘을 이용하여 잡음감쇠 성능을 개선하였다. 입력신호를 단시간 구간별로 웨이블릿 변환한 다음 1024-1024-512-neuron FNN 딥러닝 모델을 이용하여 잡음이 포함된 단일입력 음성신호로부터 잡음을 제거한다. 이는 시간영역 음성신호를 잡음특성이 잘 표현되도록 시간-주파수영역으로 변환하고 변환 파라미터에 대해 순수 음성신호의 변환 파라미터를 이용한 지도학습을 통하여 잡음환경에서 효과적으로 음성을 예측한다. 본 연구에서 제안한 잡음감쇠시스템의 성능을 검증하기 위하여 Tensorflow와 Keras 라이브러리를 사용한 시뮬레이션 프로그램을 작성하고 모의실험을 수행하였다. 실험 결과, 제안한 심층학습 알고리즘을 사용하면 기존의 적응필터를 사용하는 경우보다 30%, STFT(: Short-Time Fourier Transform) 변환을 사용하는 경우보다는 20%의 평균자승오차(MSE: Mean Square Error) 개선효과를 얻을 수 있었다.

Keywords

Acknowledgement

이 논문은 2023년도 남서울대학교 학술연구비 지원에 의해 연구되었음.

References

  1. H. Lee, "Nonlinear noise attenuator by adaptive Wiener filter with neural network," J. of the Korea Institute of Electronic Communication Sciences, vol. 18, no. 1, 2023, pp. 71-76.
  2. S. F. Boll and D. C. Pulsipher, "Suppression of acoustic noise in speech using two microphone adaptive noise cancellation," IEEE Trans. Acoust., Speech, Signal Processing, vol. ASSP-28, no. 6, Dec. 1989, pp. 752-753. https://doi.org/10.1109/TASSP.1980.1163472
  3. W. A. Harrison, J. S. Lim, and E. Singer, "A new application of adaptive noise cancellation," IEEE Trans. Acoust., Speech, Signal Processing, vol. ASSP-34, Feb. 1986, pp. 21-27. https://doi.org/10.1109/TASSP.1986.1164777
  4. S. Mallat and W. L. Hwang, "Singularity detection and processing with wavelets," IEEE Trans. on Information Theory, vol. 38, no. 2, 1992, pp. 617-643. https://doi.org/10.1109/18.119727
  5. M. J. Shensa, "The Discrete Wavelet Transform: Wedding the A Trous and Mallat Algorithms," IEEE Trans. on Signal Processing, vol. 40, no. 10, 1992, pp. 2464-2482. https://doi.org/10.1109/78.157290
  6. D. L. Donoho, "De-Noising by Soft-Thresholding," IEEE Trans. on Information Theory, vol. 41, no. 3, 1995, pp. 613-627. https://doi.org/10.1109/18.382009
  7. I. Daubechies, "The Wavelet Transform Time-Frequency Localization and Signal Analysis," IEEE Trans. on Information Theory, vol. 36, no. 5, 1990, pp. 961-1005. https://doi.org/10.1109/18.57199
  8. H. Lee, "Optimization of the number of filter in CNN noise attenuator," J. of the Korea Institute of Electronic Communication Sciences, vol. 16, no. 4, 2021, pp. 625-632.
  9. D. Rumelhart, G. Hinton, and R. Williams, "Learning representations by back-propagating errors," Cognitive modeling, vol. 5, 1988, pp. 3.
  10. J. Schmidhuber, "Deep learning in neural networks: An overview," Neural Networks, vol. 61, 2015, pp. 85-117. https://doi.org/10.1016/j.neunet.2014.09.003
  11. Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, "Gradient-Based Learning Applied to Document Recognition," Proceedings of the IEEE, vol. 86, no. 11, Nov. 1998, pp. 2278-2324. https://doi.org/10.1109/5.726791