[1]GUO Xiao-xia,YANG Hui-zhong.An improved compromise for soft/hard thresholds in wavelet denoising[J].CAAI Transactions on Intelligent Systems,2008,3(3):222-225.
Copy
An improved compromise for soft/hard thresholds in wavelet denoising
CAAI Transactions on Intelligent Systems[ISSN 1673-4785/CN 23-1538/TP] Volume:
3
Number of periods:
2008 3
Page number:
222-225
Column:
学术论文—人工智能基础
Public date:
2008-06-25
- Title:
- An improved compromise for soft/hard thresholds in wavelet denoising
- Keywords:
- wavelet transform; threshold denoising; threshold function; MSE
- CLC:
- TN911
- DOI:
- -
- Abstract:
- The characteristics of two traditional wavelet comp romise methods, referred to as the soft2thresholding method and the hard2thresholding method, were analyzed in the wavelet domain. A new comp romise thresholding method that imp roves performance is p resented, based on the standard comp romise thresholdmethod, yetmore flex2 ible and easier to treatmathematically. Compared with traditionalwavelet comp romise methods, the new app roach avoids the discontinuity of the hard2threshold method and also decreases the fixed bias between the estimated wave2 let coefficients and the decomposed wavelet coefficients of the soft2threshold method. Simulation results show that the imp roved method is better at denoising than the soft2threshold method, the hard2threshold method, the standard comp romised threshold method, aswell as other threshold function methods.
- References:
-
[1] 付炜,许山川. 一种改进的小波域阈值去噪算法[J]传感技术学报, 2006,(02) . [
2] 张维强,宋国乡. 基于一种新的阈值函数的小波域信号去噪[J]西安电子科技大学学报, 2004,(02) .
?[3] Donoho D L. De-noising by soft-thresholding .IEEE Trans. On IT. 1995,413, 41(3) :613-627 .
?[4] Donoho D L, Johnstone I M. Adapting to unknown smoothness via wavelet shrinkage [J] .Journal of American Stat Assoc, 1995,12, 12 (90) :1200-1224 .
[5] DONG Yongsheng,YI Xuming. Wavelet de-noising based on four improved functions for threshold estimation[J] .Journal of Math, 2006,26, 26 (5) :473-477 .
- Similar References:
Memo
-
Last Update:
2009-05-14