How do I deal with brightness rescaling after FFT'ing and spatially filtering images? -

louise here. i've started experimenting fourier transforming images , spatially filtering them. example, here's 1 of fireplace, high-pass filtered remove above ten cycles per image:

http://imgur.com/eca306n,nbqtmsk,ngo8eey#0 - first image (sorry, can't post images on stack overflow because haven't got enough reputation).

as can see, image dark. however, if rescale [0,1] get

http://imgur.com/eca306n,nbqtmsk,ngo8eey#0 - second image

and if raise in image power of -0.5 (we can't raise positive powers image data between 0 , 1, , smaller), this:

same link - third image

my question is: how should deal reductions in dynamic range due hi/low pass filtering? i've seen lots of filtered images online , seemed have similar brightness profiles original image, without manipulation.

should leaving centre pixel of frequency domain (the dc value) alone, , not removing when low-pass filtering?

is there commonplace transform (like histogram equalisation) should using after filtering?

or should interpret brightness reduction normal, because of information in image has been removed?

thanks advice :) best, louise

i agree connor, best way preserve brightness keep origin (dc) value unchanged. common practise. way similar image second image, because not change average gray level of image. removing using high-pass filtering set value 0 , scaling needed afterwards.