mcmc
04-26-2011, 08:55 AM
Hi guys,
Just been reading through the book and have successfully got the FFT function rlft3 working. I am trying to work out how to extend the cnvlv function to 2d, so that I can apply it to 2 images rather than signals (i.e. OriginalImage = inverse_FFT[ FFT(ObservedImage) / FFT(PointSpreadFunction) ]. The point here is that I need to find the original image from the image input to my program, which has already been blurred by a known PSF.
I am a bit unsure how I can use cnvlv to do this - I understand how it works mathematically, I just don't see a simply way to extend/use cnvlv for this (unless there isn't one...?).
Help is very much appreciated :)
Just been reading through the book and have successfully got the FFT function rlft3 working. I am trying to work out how to extend the cnvlv function to 2d, so that I can apply it to 2 images rather than signals (i.e. OriginalImage = inverse_FFT[ FFT(ObservedImage) / FFT(PointSpreadFunction) ]. The point here is that I need to find the original image from the image input to my program, which has already been blurred by a known PSF.
I am a bit unsure how I can use cnvlv to do this - I understand how it works mathematically, I just don't see a simply way to extend/use cnvlv for this (unless there isn't one...?).
Help is very much appreciated :)