f=iwfbt(c,info); f=iwfbt(c,wt,Ls);
| c | Coefficients stored in a cell-array. |
| info, wt | Transform parameters struct/Wavelet Filterbank tree |
| Ls | Length of the reconstructed signal. |
| f | Reconstructed data. |
f = iwfbt(c,info) reconstructs signal f from the coefficients c using parameters from info struct. both returned by wfbt function.
f = iwfbt(c,wt,Ls) reconstructs signal f from the coefficients c using filterbank tree defined by wt. Plese see wfbt function for possible formats of wt. The Ls parameter is mandatory due to the ambiguity of reconstruction lengths introduced by the subsampling operation and by boundary treatment methods. Note that the same flag as in the wfbt function have to be used, otherwise perfect reconstruction cannot be obtained. Please see help for wfbt for description of the flags.
A simple example showing perfect reconstruction using idtwfb:
f = gspi;
J = 7;
wt = {'db6',J};
c = wfbt(f,wt);
fhat = iwfbt(c,wt,length(f));
% The following should give (almost) zero
norm(f-fhat)