感觉很好的课题，但实际上很难，理论上signals are not always smooth, and generally with many singularities, the relativity over singularities is week, so the wavelet transforms with large supporting length seem not fit for. But using adaptive algorithm must relate with quantity process. I think the coefficients at the singularities must big, so will be significant in zerotree, am i correct?
浙公网安备 33010602011771号