Performance Analysis of a Robust Wavelet Threshold for Heavy-Tailed Noises

Abstract:

The interesting signal is often contaminated by heavy-tailed noise that has more outliers than Gaussian noise. A robust wavelet threshold based on the minimax description length principle is derived in the ε-contaminated normal family for maximizing the entropy. Compared with classical threshold based on Gaussian assumption, the robust threshold can eliminate the heavy-tailed noise better, even if the precise value of ε is unknown, which shows its robustness. The further experiment shows that soft threshold is more suitable than hard threshold for robust wavelet threshold technique.

• Periodical: Applied Mechanics and Materials (Volumes 40-41)
• Pages: 979-984
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.40-41.979
• Online since: November 2010
• Authors: Guang Fen Wei, Feng Su, Tao Jian
• Keywords: Hard Threshold, Heavy-Tailed Noise, Robust Wavelet Threshold, Signal Detection, Soft Threshold

References

[1] G.F. Wei , T. Jian, R.J. Qiu: Proc. 8th Inter. Conf. on Signal Processing Vol. 1(2006), p.53 IEEE Press.

[2] T. Jian, Y. He, F. Su: Signal Processing Vol. 23 (2007), p.79 (in Chinese).

[3] S. Mallat, W.L. Hwang: IEEE Trans. on Information Theory Vol. 38 (1992), p.617.

[4] D.L. Donoho: IEEE Trans. on Information Theory Vol. 41(1995), p.613.

[5] T. Jian, F. Su, Y. He: J. of Electro. & Infor. Technology Vol. 29 (2007), p.1864 (in Chinese).

[6] H. Krim, I. C. Schick: IEEE Trans. on Information Theory Vol. 45(1999), p.898.

[7] P. Huber: Ann. Math. Stat Vol. 35(1964), p.1753.

[8] I. Djurovic, V. Katkovnik, L. Stankovic: IEEE Inter. Conf. on Acoustics, Speech, and Signal Processing Vol. 6 (2001), p.3517.

[9] T. Jian, F. Su, D.F. Ping: Modern Radar Vol. 29 (2007), p.55 (in Chinese).