Calculating the Minimum and The Maximum of a Slidig Window

Hello,

How would you create the fastest implementation to extract the Minimum and the Maximum of a Local Sliding Window over an image?

So far I used this:

localMaxImage = colfilt(inputImage, [winLength winLength], 'sliding', @max);
localMinImage = colfilt(inputImage, [winLength winLength], 'sliding', @min);

Yet it is really slow. I know I can use `imerode` and `imdilate`, Yet I'm looking for a toolbox free implementation. Thank You.

[José-Luis]

Royi, you have asked 23 questions in this forum so far and not accepted a single one. Have you had bad luck with your answers? If not, seeing that might make contributors here less inclined to help you.

[Royi Avital]

Well, I guess as you see, I don't remember getting the right answer. You may look at my profile in StackExchange, I accept more than 80% of answers.

Anyhow, When I try answering someone, I never look at that. We're not trying to get scored, we're trying to assist people and build well organized information source.

Answer 1

[Matt J]

I've never used it, but this FEX contribution is MEX-driven, and so should be fast

http://www.mathworks.com/matlabcentral/fileexchange/24705-min-max-filter

Answer 2

[Anand]

What do you mean toolbox-free? colfilt is also from the Image Processing Toolbox. Can you use ordfilt2

localMaxImage = ordfilt2(inputImage, winLength^2, true(winLength));
localMinImage = ordfilt2(inputImage, 1          , true(winLength));

This should be faster than colfilt, but maybe you can't use it.

[Royi Avital]

I think it is also from the image toolbox.