How does colfilt work ? Is there any mathematical explanation how it work ?

Question

Parag 2017-1-9

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/319727-how-does-colfilt-work-is-there-any-mathematical-explanation-how-it-work

评论： Image Analyst 2017-1-11

Hi Guys,

I wanted to understand the working of colfilt. When we provide the block size and function (ex.colfilt(img,[3 3],'fun');) does it make overlapping boxes if not how does the output image is of same size. If there is any document wherer I can find the mathematical explanation ?.

I will really appreciate for your help.

Thanks

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Image Analyst 2017-1-10

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/319727-how-does-colfilt-work-is-there-any-mathematical-explanation-how-it-work#answer_250019

I agree it's confusing, so maybe that's why I've never found a need to use it. Supposedly it can make processing faster but since typical operations I do on an image take like a second or less, time has never been something I've worried about.

It looks like it has a 'distinct' (jumping by several pixels) option and a 'sliding' (move by 1 pixel) option.

2 个评论
显示无隐藏无

Parag 2017-1-10

Thank you for your comment. Is Colfilt use a convolution-based method to apply specific function ? .For example I am applying a fun x to image I(x,y) then the out put image will be I(x,y)*fun is that right. Thanks

Image Analyst 2017-1-11

No, it doesn't do convolution. imfilter() is more like that (except imfilter doesn't flip the kernel like conv2 does). colorfilt() applies the function you supply, so it's more like nlfilter().

请先登录，再进行评论。