Odd output from ind2sub

I'm trying to spit out indices for max values in a large matrix at a specific point in time (in this case obs 1560, t(:, :, 1560))

>> [I J]=ind2sub(size(t), (max(max(t(:, :, 1560)))))

But my I is equal to the actual value in the cell, not the row

I =

     28.5048

And J makes no sense:

J =

       1

Is the error in my code obvious to anyone? Thanks,

Answer 1

[Andrei Bobrov]

s = size(t);
[val,ij] = max(reshape(t(:, :, 1560),[],1);
[ii,jj] = ind2sub(s(1:2),ij);
idxs = [ii,jj,1560];

or

ind = prod(s(1:2))*1559 + ij;

[Kate]

Thanks Andrei, reshape was the addition that I needed.

Answer 2

[dpb]

(max(max(t(:, :, 1560))))

is a single value for the maximum of the values of the requested plane. Thus there isn't any actual location to be converted from being returned.

If you write

[i,j]=ind2sub(size(t), find(max(max(t(:, :, 1560)))))

instead, you'll always get [1,1] because the result of the double max() operation is scalar.