基于局部余弦变换的多普勒信号降噪方法研究

Doppler signal de-noising based on local cosine transform

提出一种基于自适应局部余弦变换和非负Garrote取阈值的多普勒超声信号降噪方法.首先,对信号进行局部余弦变换,并且搜索描述信号的最优基;然后,利用非负Garrote方法对最优基系数取阈值;最后,由取阈值后的系数进行重构得到真实信号的估计.仿真结果表明,该方法在低信噪比情况下优于基于时移不变小波变换的降噪方法,是一种可行而且有效的降噪方法.

A novel approach based on the adapted local cosine transform and the non-negative garrote thresholding method was proposed to remove noise from Doppler ultrasound signals. At first, the local cosine transform was performed on the signal of interest, and was followed by a best basis searching algorithm to best match the signal. Then a non-negative garrote thresholding method was used to shrink the coefficients, from which the de-noised signal was reconstructed. The simulation results show that this approach is superior to the one based on the translation-invariant wavelet transform, especially under low SNR conditions. Therefore this approach is a practical and powerful tool for de-noising the Doppler ultrasound signal.