How can I get the number between 2 numbers?

I have 2 arrays and the index of the stable numbers(n)

n=[1 2 7 9];
PA = [1 3 10 1 2 8 6 2 7 11 7 9 5 9];
PB = [1 10 1 2 11 6 2 7 3 7 9 5 8 9];

I want to store those numbers between the couple same numbers( 1 1, 2 2, 7 7, 9 9) separately in the cell like below.

CA={[3 10];[8 6];[11];[5]};
CB={[10];[11 6];[3];[5 8]};

Answer 1

[dpb]

ix=cell2mat(arrayfun(@(n)find(PA==n),n,'uni',0).');
PA=arrayfun(@(i1,i2)PA(i1+1:i2-1),ix(:,1),ix(:,2),'uni',0);

PB should be obvious... :)

Not using two named variables if the number elements is always the same could let write more generic code...or a cell array for them as well if different. If the A, B suffixes are significant, the row or cell index would be the identifier.

[Hang Vu]

Thank you so much for the help!

Answer 2

[Image Analyst]

If you have the Image Processing Toolbox, you can simply use ismember() followed by regionprops() and in 2 lines of code be done. Try this:

% Define input data:
n=[1 2 7 9];
PA = [1 3 10 1 2 8 6 2 7 11 7 9 5 9];
PB = [1 10 1 2 11 6 2 7 3 7 9 5 8 9];

% Find out where n occurs in PA:
[ia, ib] = ismember(PA, n)
% Extract the values for the regions into a structure array.
props = regionprops(~ib, PA, 'PixelValues');

% Basically we're already done, but if you want the values
% in a cell array instead of a structure array, do this:
for k = 1 : length(props)
	CA{k} = props(k).PixelValues;
end
% Display what we got in the command window.
celldisp(CA)

You get exactly what you asked for. Repeat for CB. Personally I prefer structure arrays over cell arrays but you can use whichever you want.

[dpb]

"...I prefer structure arrays over cell arrays ..."

Man, I can NEVER figure out how to end up dereferencing stuff once I get it into a structure, array, IA... :( I always seem to end up in a conundrum of subscripting expressions that just never work or "you can't get there from here!" problems. :(

The doc is, to me at least, opaque at best and the examples are always trivial that never seem to address "real world".

But, that's just me...I was so frustrated so much early on I basically quit even trying...

[Image Analyst]

You simply use the index like a regular index, and the field are whatever names they are or you want to call them. For example, if you also had measurements of the indexes in the region, the average value in the regions, and the length of the regions you could simply use the same index and the proper field names, like props(k).PixelIdxList, props(k).MeanIntensity, props(k).Area. props just comes out with whatever fields you asked for when you called regionprops() so they're there automatically. You can also have the measurements returned in a table, which is also more convenient than a cell array if you have multiple regions in your data.

[Hang Vu]

Image Analyst, Thank you so much for your time! It's really cool.