| [ KAIST ] in KIDS 글 쓴 이(By): iLUSiON (화려한집념) 날 짜 (Date): 2003년 3월 5일 수요일 오전 04시 51분 46초 제 목(Title): Re: [Q]필터 잘아시는분 oh oh~~~ darkman's additional info clears the problem. this is an ill-posed problem. exactly the same problem puzzed the one student doing fMRI analysis and asked me about an advice last term. :) forget about trying to get the exact solution. the best you will do is getting the least squres estimation of the original signal. which is the case for most of all convolution and deconvolution of signals. one numerical trick for the solution not to have spikes or behaving irratically is to make signal as smooth as possible by nonparemetric smoothing. then you define the derivative of signal y(t) as follwing Dy(t) = d K*y(t)/dt the derivative of nonparametric smoothing of signal y(t). you can show the asymptotic convergence of this definition to the standard derivative. ok.... i will stop short of today's lecture. i have to play MORROWiND. i am hooked. ^^ iLUSiON whitepolarcow@hotmail.com |