论文部分内容阅读
This talk proposes two kinds of such analytic approximations of which one is called maximal-energy AFDs,including core AFD,Unwending AFD and Cyclic AFD; and the other is again linear combinations of Szeg\"o kernels but generated through SVM methods.The proposed methods are based on the fact that the imaginary part of an analytic signal is the Hilbert transform of its real part.This paper for the first time promotes the complex analytic method for computing Hilbert transforms.Experiments show that such computational methods are as effective as the commonly used one based on FFT.