Closed MathiasHurst closed 2 years ago
Hi Tomas, are there any news on this? Bests, Mathias
Not yet, but thanks for reminding me. Next week I should have some time to check this.
Hi Tomas, sorry meanwhile I used the python version of this and forgot about the issue. Filter for weakly absorbing sample with a slowly variing phase with homogenous composition is:
CTF = 2 sin(pi lambda z_i k^2) - 2 beta/delta cos(pi lambda z_i * k^2)
FFT(PHI) = SUM_i [ ( FFT(I_i -1) ) * CTF ] / SUM_i [ CTF^2 + epsilon ]
There are some variiations for pure phase or non homogenous objects, but I think this is the approach with the best ratio between applicabilitiy and simplictiy. If used with real data, Image alignment is really crucial.
Can you implement the following phase retrieval filter, for multidistance phaseretrieval? Usually 2-4 distances will be used.
1/ ( SUM_i( 2 sin(pi lambda z_i k^2) ) + 10^-regularization_rate )
with frequency k propagation distance z wavelength lambda