Convoluting simulated image with real detector MTF

Hi everyone,

I would like to include the effect of the detector response on transmission electron microscope images I am simulating. In principle, I understand that the simulated image has to be convolved with the detector modulation transfer function. To practically attain this I would have to firstly FFT the image and then multiply it by the MTF. After this I would have to perform an iFFT of the result to get an image in real space which has the MTF influence.

I am struggling to understand how I can convolve a 2D image with a 1D detector MTF.

Examples of the MTF and images I am simulating are attached.

[Philip Plummer]

A little late to help you probably, but I would just make assumptions regarding the MTF, specifically that it was symmetric with regards to both spatial frequency dimensions (fx and fy for discussion). I would then utilize Matlab's meshgrid function and generate a 2-D MTF to apply to my image.

[Arthur Moya]

@Philip Plummer,

No you are not late at all. I have a code that was given to me a while back but I don't think I totally appreciate the implications of using it. I was confused by the way it was written and so I am not confident using it without totally grasping what it is doing. You are correct that the MTF is symmetric with regards to y and x. With meshgrid I get a 2D display of the MTF which is great. However, I failed to clarify that the x values of the MTF have to be rotationally repeated along y and centred at zero. What I would like to do after that would be centre and pad the output (725x725) with zeros to create a 2kx2k double matrix which I can then convolve with the FFT of the image.

Thanks for replying.