what (:, :, :) syntax does?

Hi, I'm new in MatLab and my teacher gave us a code to crop images and split them into his RGB components. This section it's suposed to crop the image (I1) I think (am I right?) and display the cropped image (I2)

u=round(ginput(2));
    P12 = u(1,2);
    P22 = u(2,2);
    P11 = u(1,1);
    P21 = u(2,1);
    
    I2=I1(P12:P22,P11:P21,:);
    figure
    imshow(I2)
    axis equal

I think I know what the first line does, it gets two inputs from the mouse and then round it to the nearest integer to finally save them into a vector 'u' right?, in that case, my doubt is what those u(1,2) means, how do they work? what are those 1 and 2 into them? And, how that syntax I2=I1(P12:P22,P11:P21,:) works? why is only the : in the final?

I know I could get the answer by reading the documentation, but what documentation do I have to read? Thank you in advance.

[Lateef Adewale Kareem]

you can use that to index 3 dimensional space in a 4 or higher dimensional data.

The question is why would you need that in the first place

[Voss]

@Emmanuel Arias Polanco: Here is some documentation you can read that will answer all of those questions:

https://www.mathworks.com/company/newsletters/articles/matrix-indexing-in-matlab.html

Answer 1

[Jonas]

u=round(ginput(2));

gets two points with rounded x and y values. the values are stored as [x1 y1; x2 y2]

it follows that

P12 = u(1,2); % first y coordinate
    P22 = u(2,2); % second y coordinate
    P11 = u(1,1); % first x coordinate
    P21 = u(2,1); % second x coordinate

following

I1(firstY:secondY,firstX:secondX,:)

which cuts in horizontal and vertical between the specified x and y values.

the : in the third dimension means that all entries along the third dimension are kept. the object has three dimensions since it has 3 layers per point, containing the rgb values of the colored pixel

so the given code just cuts the image but does not split into the rgb contents, since all entries are kept along the 3rd dimension.

if you want to see the split components, e.g. to see only the red channel also displayed in red, use e.g.

I3=I2; % crrate copy of the cut image part
I3(:,:,2:3)=0; % set all values of the g and b layer to 0
imshow(I3);