For which type of noise kuwahara filter is most affective? ( inscale of PSNR etc)..

The kuwahara filter which comes under in nonlinear filtering section which having a kuwahara operator of windows 5,9 etc which can be used for the image noise reduction but i want to know that for which type of noise does kuwahara filter is most effective with and if possible the effective windows size also.

Answer 1

[Image Analyst]

I don't know what type of noise it's most effective on. It seems to be an ad hoc algorithm that was developed without assuming any particular type or distribution of noise.

The "best" window size depends on how much smoothing and change to the picture you're willing to put up with. Biggest windows provide more smoothing and greater departure from the original image.

The filter can be done by using imfilter() or conv2(), and with stdfilt(), and then combining outputs of those.

[SARATH JV]

Thanks for the kind reply

[Image Analyst]

For what it's worth, I'm attaching my demo for a gray scale Kuwahara filter.