ImageJ max and min filters

I have been using the max and min filters in ImageJ to erode and dialate images.

Now I am trying to replicate this functionality in MATLAB.

Using Square matrixes for the neighborhood, the imerode and imdialate results are not as "smooth" as what ImageJ produces.

What extra technique is ImageJ using to produce "smooth" erosions?

[Adam]

Sounds more like a question for an ImageJ forum. Have you asked there?

[Dante Basile]

Yes, I have, I was just hoping to come across someone with experience replicating ImageJ operations in MATLAB

Answer 1

[Dante Basile]

This is the function I have come up with for replicating the masks used by ImageJ

function[kMask] = ijRadMask(radius)
% IJRADMASK get mask for imerode/imdilate which is identical to the one used by imageJ for erosion/dilation
%   input: radius in pixels to be eroded/dilated
%   output: disk mask for erosion/dilation
    r2 = floor(radius^2 + 1); %simulate java int conversion
    kRadius = int32(floor(sqrt(r2 + 1e-10))); 
    kernel = int32(zeros(kRadius + 1, 1)); 
    kernel(kRadius + 1) = kRadius; %set kernel center
    for y = 1:double(kRadius) %make kernel
        dx = int32(floor(sqrt(r2 - y^2 + 1e-10)));
        kernel((kRadius + 1) - y) = dx;
    end
    kMask = zeros(kRadius*2 + 1); %mask for filter
    kHeight = 2*kRadius + 2;
    mCenter = kRadius + 1;
    kMask(mCenter, 1:size(kMask, 2)) = 1; %horizontal line through center
    for i = 1:(size(kernel, 1) - 1)
        kMask(i, (mCenter - kernel(i)):mCenter + kernel(i)) = 1;
        kMask(kHeight - i, (mCenter - kernel(i)):mCenter + kernel(i)) = 1;
    end
end