In Matlab I would like to do a sort on color values in a particular order.
Say for instance, Sort by Hue, then Sat, then Intensity (V). Or, by V, H then S.
Something like this:
% UNSORTED
A(:,:,1) = [1 1 1 1 1 2 2 2 1 1 1];
A(:,:,2) = [1 1 1 1 1 3 2 2 3 2 2];
A(:,:,3) = [3 1 4 3 2 1 2 1 1 2 1];
% RESULT
A(:,:,1) = [1 1 1 1 1 1 1 1 2 2 2]
A(:,:,2) = [1 1 1 1 1 2 2 3 2 2 3]
A(:,:,3) = [1 2 3 3 4 1 2 1 1 2 1]
I've struggled a day on this with no luck. Any help?
One method could be:
% UNSORTED
A(:,:,1) = [1 1 1 1 1 2 2 2 1 1 1];
A(:,:,2) = [1 1 1 1 1 3 2 2 3 2 2];
A(:,:,3) = [3 1 4 3 2 1 2 1 1 2 1];
for ii = 1:size(A,1)
A(ii,:,:) = reshape(sortrows(squeeze(A(ii,:,:)),[1:3]),[1,size(A,2),size(A,3)])
end
So we extract each 2D matrix, we use squeeze to delete the singleton dimension, then we use sortrows to sort each rows (with this priority 1->2->3). And finally we reshape this 2D matrix to restore the singleton dimension and obtain a 3D matrix.
Related
So i have two very long matrices. A sample is given below:
First_Matrix:
A = [...
1 1 1;
1 1 2;
1 1 3;
1 2 1;
1 2 2;
1 2 3;
1 3 1;
1 3 2;
1 3 3];
Second Matrix
B = [...
1 1 916;
1 2 653;
1 3 114];
And I would like a thirds matrix that would combine the first matrix with the third column of the second matrix, based on the values in the first two column of the 2 matrices matching (being the same).
So Ouput_Matrix:
C = [...
1 1 1 916;
1 1 2 916;
1 1 3 916;
1 2 1 653;
1 2 2 653;
1 2 3 653;
1 3 1 114;
1 3 2 114;
1 3 3 11];
What would be the best way to do this?
Thanks in advance
Use the second output of ismember with the 'rows' option to get the indices of the matching, from which you can easily build the result:
[~, ind] = ismember(A(:, [1 2]), B(:, [1 2]), 'rows');
C = [A B(ind, 3)];
The for loop isn't pretty and might slow you down if B is very long. But I don't think it's possible to avoid (edit: it seems it is).
A = [1 1 1
1 1 2
1 1 3
1 2 1
1 2 2
1 2 3
1 3 1
1 3 2
1 3 3];
B = [1 1 916
1 2 653
1 3 114];
C = [A zeros(size(A,1),1)];
for i = 1:size(B,1)
C(all(B(i,1:2)==A(:,1:2),2),4) = B(i,3);
end
C =
1 1 1 916
1 1 2 916
1 1 3 916
1 2 1 653
1 2 2 653
1 2 3 653
1 3 1 114
1 3 2 114
1 3 3 114
It is possible to achieve what you want without a for loop, but it may not be the most optimal implementation:
n = size(B, 1); % number of rows in B
B_(1, :, :) = B'; % convert to 3D matrix to be able to use elementwise comparision
x = squeeze(all(bsxfun(#eq, A(:, 1:2), B_(1, 1:2,:)), 2)); % x(i, j) == 1 if row A(i, :) matches B(j, :)
index = x * (1:n)'; % row B(index(i), :) corresponds with row A(i, :)
A(:, 4) = B(index, 3); % add data to A
An alternative formulation for x, without conversion to 3D, is:
x = bsxfun(#eq, A(:, 1), B(:,1)') & bsxfun(#eq, A(:, 2), B(:,2)');
The disadvantage of this method is that it less extendable to more matching columns.
Creating a number pattern based on the length of array, the example is small just showing what it would look like if the array had a size of 5,6 or 7. (the actually length of arrays will be around 400,000)
How can I go about getting this numerical pattern. I was thinking it looked a little like Pascal's Triangle but the numbers are off.
I'm using Octave 3.8.1 which is like Matlab
example: If an array has the length of 5 the series would be
a1=[1 1 1;1 2 2;1 2 3;1 2 2;1 1 1]
a2=[1 1 1;2 2 1;3 2 1;2 2 1;1 1 1]
example: If an array has the length of 6 the series would be
a1=[1 1 1;1 2 2;1 2 3;1 2 3;1 2 2;1 1 1]
a2=[1 1 1;2 2 1;3 2 1;3 2 1;2 2 1;1 1 1]
example: If an array has the length of 7 the series would be
a1=[1 1 1 1;1 2 2 2;1 2 3 3;1 2 3 4;1 2 3 3;1 2 2 2;1 1 1 1]
a2=[1 1 1 1;2 2 2 1;3 3 2 1;4 3 2 1;3 3 2 1;2 2 2 1;1 1 1 1]
See image below:
This should be quite fast:
n = 9;
a1 = bsxfun(#min, min(1:n,n:-1:1).', 1:ceil(n/2));
a2 = a1(:,end:-1:1);
This does what you want:
n = 7;
a1 = cumsum(tril(ones(ceil(n/2))), 2);
a1 = a1([1:end end-mod(n,2):-1:1],:);
a2 = fliplr(a1);
A slighty different approach:
For m = ceil(n/2); the bsxfun can be substituted by
X = gallery('minij',m);
which is equal to:
X = bsxfun(#min, 1:m,(1:m).')
so you get a single block and you just need to concatenate the output.
a1 = [X; flipud(X)]
if n ~= 2*m; a1(m,:) = []; end %// clear one row if necessary
a2 = fliplr(a1)
for n = 6 you get
a1 =
1 1 1
1 2 2
1 2 3
1 2 3
1 2 2
1 1 1
and a2 accordingly mirrored. and for n = 5
a1 =
1 1 1
1 2 2
1 2 3
1 2 2
1 1 1
given matrix A of size 6 by 6 contain blocks of numbers,each block of size 2 by 2, and outher reference matrix R of size 2 by 12 also contain blocks of numbers, each block of size 2 by 2. the perpse of the whole process is to form a new matrix, called the Index matrix, contain index's that refer to the position of the blocks within the matrix A based on the order of the blocks within the reference matrix R. and here is an exemple
matrix A:
A =[1 1 2 2 3 3;
1 1 2 2 3 3;
1 1 3 3 4 4;
1 1 3 3 4 4;
4 4 5 5 6 6;
4 4 5 5 6 6 ]
matrix R:
R=[1 1 2 2 3 3 4 4 5 5 6 6;
1 1 2 2 3 3 4 4 5 5 6 6 ]
the new matrix is:
Index =[1 2 3;
1 3 4;
4 5 6]
any ideas ?
With my favourite three guys - bsxfun, permute, reshape for an efficient and generic solution -
blksz = 2; %// blocksize
num_rowblksA = size(A,1)/blksz; %// number of blocks along rows in A
%// Create blksz x blksz sized blocks for A and B
A1 = reshape(permute(reshape(A,blksz,num_rowblksA,[]),[1 3 2]),blksz^2,[])
R1 = reshape(R,blksz^2,1,[])
%// Find the matches with "bsxfun(#eq" and corresponding indices
[valid,idx] = max(all(bsxfun(#eq,A1,R1),1),[],3)
%// Or with PDIST2:
%// [valid,idx] = max(pdist2(A1.',reshape(R,blksz^2,[]).')==0,[],2)
idx(~valid) = 0
%// Reshape the indices to the shapes of blocked shapes in A
Index = reshape(idx,[],num_rowblksA).'
Sample run with more random inputs -
>> A
A =
2 1 1 2
1 2 2 1
1 1 1 1
2 2 2 2
1 2 2 1
1 2 1 1
>> R
R =
2 1 1 1 1 2 2 2 1 1 1 1
2 1 2 1 1 2 2 1 2 2 2 1
>> Index
Index =
0 0
5 5
3 0
Given A is symmetry matrix with size n and
A =
1 2 3 4 5 % The Position
1 [0 5 2 4 1
2 5 0 3 0 2
3 2 3 0 0 0
4 4 0 0 0 5
5 1 2 0 5 0]
B is a row vector that permute the matrix A row and column
B = [2 4 1 5 3]
The output that I want is
C =
2 4 1 5 3 % The New Position given by Matrix B
2 [0 0 5 2 3
4 0 0 4 5 0
1 5 4 0 1 2
5 2 5 1 0 0
3 3 0 2 0 0]
I can get the output by using simple for loop
index = [2,4,1,5,3];
C = zeros(5,5);
for i = 1:5
for j = 1:5
% Position of in square matrix n
% (i,j) = (i-1)*n + j
C(i,j) = A((index(i)-1)*5+index(j));
end
end
However, if I want to permute a matrix with size 80x80, then I need to run 1600 times in order to get the output.
Is there any simple trick to do it instead of using for loop?
You should be able to rearrange your matrices as follows:
C = A(index,index);
This rearranges each dimension according to the index variable independently.
i'm trying to find local maxima of a vector of numbers using MATLAB. The built-in findpeaks function will work for a vector such as:
[0 1 2 3 2 1 1 2 3 2 1 0]
where the peaks (each of the 3's) only occupy one position in the vector, but if I have a vector like:
[0 1 2 3 3 2 1 1 2 3 2 1 0]
the first 'peak' occupies two positions in the vector and the findpeaks function won't pick it up.
Is there a nice way to write a maxima-finding function which will detect these sort of peaks?
You can use the REGIONALMAX function from the Image Processing Toolbox:
>> x = [0 1 2 3 3 2 1 1 2 3 2 1 0]
x =
0 1 2 3 3 2 1 1 2 3 2 1 0
>> idx = imregionalmax(x)
idx =
0 0 0 1 1 0 0 0 0 1 0 0 0
Something much easier:
a = [1 2 4 5 5 3 2];
b = find(a == max(a(:)));
output:
b = [4,5]
a = [ 0 1 2 3 3 2 1 2 3 2 1 ];
sizeA = length(a);
result = max(a);
for i=1:sizeA,
if a(i) == result(1)
result(length(result) + 1) = i;
end
end
result contains the max, followed by all the values locations that are equal to max.