freeburma

matlab

You need MATLABŪ to run this scripts by your own.

duck_1

Original image Duck.png

duck_2

Same image but grayscaled.

duck_3

Gaussian filter appended

Download script1.m
Invoke with:

Invoke with:

script1('Duck.png')

duck_4

Unsharp masking operation

Download script2.m
Invoke with:

script2('Duck.png')

duck_5

Sharp filtered

Download script3.m
Invoke with:

script3('Duck.png')

periodical interferences

Periodical interferences occure, if some frequences are overlayed by other frequences. Those frequences can be good spotted by looking at the fourier transformed image. Dark regions of the fourier specturm indicates no or very low frequences, bright regions high frequences. The disturbing frequences can be removed by setting them to zero. This is achieved in the java application iFilter by painting over all bright white spots (except the centered main spot) in the fourier transformed image. An inverse fourier transformation of the modifed fourier spectrum lets you see the corrected image. The results depend on the users mouse skills. Try it by yourself!

clown_1

Nasty interferences all over a pretty clown face.

clown_2

The clown's fourrier spectrum shows the interferences (white stars).

clown_3

My feeble attempt to paint over them.

clown_4

Much better now.

Download the zip-archive that contains iFilter & 4 images to filter here.