I'm trying to copy part of a matrix (matrix 1) in matlab to another empty matrix of zeros (matrix 2) so that the section I copy from matrix 1 has the same indices in matrix 2, e.g.
Matrix 1 (mat1):
0 3 0 0 2 4 1 2 6
1 3 4 2 0 0 0 2 0
0 2 6 1 3 6 6 1 1
0 0 0 2 1 3 3 1 0
1 4 5 2 3 3 0 0 1
Matrix 2 (mat2) desired output:
0 0 0 0 0 0 0 0 0
0 0 4 2 0 0 0 0 0
0 0 6 1 3 6 6 0 0
0 0 0 2 1 3 3 0 0
0 0 0 0 0 0 0 0 0
I've tried something like
mat2([2:4],[3:7]) = mat1([2:4],[3:7])
but of course it doesn't work... any ideas of an efficient way to do this? I couldn't find another thread to help with this problem.
Thanks!
It does work. You just need to create mat2 first:
mat2 = zeros(size(mat1));
mat2(2:4, 3:7) = mat1(2:4, 3:7);
Note that you don't need the square brackets on those ranges.
Do this:
mat2 = zeros(size(mat1));
Before copying over.
Related
I have written some code that compresses a matrix to remove zero columns and rows, but I can't work out how to reconstruct the original matrix.
Say I have a matrix:
A = [ 0 3 0 2 1 0 6
3 0 0 4 8 0 5
0 0 0 0 0 0 0
2 4 0 0 2 0 1
1 8 0 2 0 0 7
0 0 0 0 0 0 0
6 5 0 1 7 0 0 ]
Here rows/columns 3 and 6 are empty, so my compression function will give the output:
A_dash = [ 0 3 2 1 6
3 0 4 8 5
2 4 0 2 1
1 8 2 0 7
6 5 1 7 0 ]
A_map = [ 1 2 4 5 7]
Where A_map is a vector mapping the indicies of the rows/columns of A_dash to A. This means that if A_map(3) = 4, then row/column 4 of A is the same as row/column 3 of A_dash - ie. a row/column of zeroes must be inserted between columns/rows 2 and 3 in A_dash
What is the easiest way people can suggest for me to recreate matrix A from A_dash, using the information in A_map?
Here is what I have got so far:
% orig_size is original number of columns/rows
c_count = size(A_dash,1);
A = zeros(c_count, orig_size); % c_count rows to avoid dimension mismatch
for ii = 1:c_count
A(:,A_map(ii)) == A_dash(:,ii);
end
This gives me the right result column-wise:
A = [ 0 3 0 2 1 0 6
3 0 0 4 8 0 5
2 4 0 0 2 0 1
1 8 0 2 0 0 7
6 5 0 1 7 0 0 ]
However, I'm not sure how i should go about inserting the rows, i suppose i could copy the first 1:i rows into one matrix, i:end rows to a second matrix and concatenate those with a zero row in between, but that feels like a bit of a
clunky solution, and probably not very efficient for large sized matrices..
Otherwise, is there a better way that people can suggest I store the map information? I was thinking instead of storing the mapping between column/row indices, that I just store the indices of the zero columns/rows and then insert columns/rows of zeros where appropriate. Would this be a better way?
You've got the indices of the valid rows/columns. Now all you've got to do is put them in a new matrix of zeros the same size as A:
B=zeros(size(A));
B(A_map,A_map)=A_dash
B =
0 3 0 2 1 0 6
3 0 0 4 8 0 5
0 0 0 0 0 0 0
2 4 0 0 2 0 1
1 8 0 2 0 0 7
0 0 0 0 0 0 0
6 5 0 1 7 0 0
Just to check...
>> A==B
ans =
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
1 1 1 1 1 1 1
A and B are equal everywhere, so we've reconstructed A.
I have a very large matrix(around 4000000x2) , and it has 1s sprinkled throughout the matrix. What I want to do is that I just want to add up all the 1s in one row.
For example, if I have a matrix like this:
A = [0 0 4 1 0 0 1
1 0 5 0 7 0 1
5 6 0 8 1 0 6
0 9 5 1 0 0 0]
Is there a way of summing up all of the 1's a row? For example, here it would be:
sum = [2
2
1
1]
I know that if you want to add up the whole row, you can use sum(A,2). But is there a way in matlab to add up all of a specific number? I'm new to matlab and I would greatly appreciate any help, thank you!!
Generate an array that has a 1 everywhere A has a 1, and 0 everywhere else:
>> A == 1
ans =
0 0 0 1 0 0 1
1 0 0 0 0 0 1
0 0 0 0 1 0 0
0 0 0 1 0 0 0
Then you can just use sum:
sum(A == 1, 2)
Let's say I have a square matrix M:
M = [0 0 0 0 0 1 9; 0 0 0 0 0 4 4; 0 0 1 1 6 1 1; 0 1 2 9 2 1 0; 2 1 8 3 2 0 0; 0 8 1 1 0 0 0; 14 2 0 1 0 0 0]
0 0 0 0 0 1 9
0 0 0 0 0 4 4
0 0 1 1 6 1 1
M = 0 1 2 9 2 1 0
2 1 8 3 2 0 0
0 8 1 1 0 0 0
14 2 0 1 0 0 0
Now I'd like to calculate two different cumulative sums: One that goes from the top of each column to the element of the column, that is a diagonal element of the matrix, and one that goes from the bottom of the column to the same diagonal element.
The resulting matrix M'should therefore be the following:
0 0 0 0 0 1 9
0 0 0 0 0 4 5
0 0 1 1 6 2 1
M' = 0 1 3 9 4 1 0
2 2 8 5 2 0 0
2 8 1 2 0 0 0
14 2 0 1 0 0 0
I hope the explanation of what I'm trying to achieve is comprehensible enough. Since my matrices are much larger than the one in this example, the calculation should be efficient as well...but so far I couldn't even figure out how to calculate it "inefficiently".
In one line using some flipping and the upper triangular function triu:
Mp = fliplr(triu(fliplr(cumsum(M)),1)) ...
+flipud(triu(cumsum(flipud(M)),1)) ...
+flipud(diag(diag(flipud(M))));
The following will do the job:
Mnew = fliplr(triu(cumsum(triu(fliplr(M)),1))) + flipud(triu(cumsum(triu(flipud(M)),1)));
Mnew = Mnew - fliplr(diag(diag(fliplr(Mnew)))) + fliplr(diag(diag(fliplr(M))));
But is it the fastest method?
I think logical indexing might get you there faster
I need to get a 1-D profile of a circular image, for example 256x256 sin(R) image
I've written a matlab function for the task but it turns out to be very un-efficient.
the function averages over radius intervals of the original images.
matlab profiler reveals that the first line in the for-loop [indxs=find(...)]
takes ~86% of the running time.
i need to run the function on a some thousands of simulated images (some larger then 256x256) and it takes very long time to complete.
does anyone knows how can i make this code run faster?
maybe someone has another, more efficient way to do the task??
i also tried to convert to function into C++ & mex file using matlab coder
but it took longer (x3) to perform the task, might be because the sub-function- "findC"
uses some 2D-ffts to find the center of the image.
Thanks you All,
Dudas
My Matlab function:
function [sig R_axis Center]= Im2Polar (imR,ch,Center_Nblock)
% Converts Circular image to 1-D sig
% based on true image values w/o interpolation
% Input -
% imR - circular sinuns image
% ch - number of data-points in output signal (sig)
% Center_Nblock - a varible related to the image center finding method
% Output -
% sig - 1D vector of the circular image profile
% R_axis - axis data-points for sig
% Center - image center in pixels
[Mr Nr] = size(imR); % size of rectangular image
[Center]=findC(imR,Center_Nblock);
Xc=Center(1);
Yc=Center(2);
rMax=sqrt((Mr/2)^2 + (Nr/2)^2);
x=[0:1:Mr-1]-Xc+1;
y=[0:1:Nr-1]-Yc+1;
[X,Y]=meshgrid(x,y);
[TH,R] = cart2pol(X,Y);
% Assembling 1-D signal
sig=single([]);
ii=1;
dr=floor(rMax)/ch;
V=dr:dr:floor(rMax);
for v=V
indxs=find((v-dr)<=R & R<v);**
sig(ii)=mean(imR(indxs));
Nvals(ii)=length(indxs);
ii=ii+1;
end %for v
R_axis=V-dr/2;
end % of function
Following from the comments here's an example of something I might try. Let's work with a 9x9 example. Suppose you have the following annulus.
A =
0 0 0 0 0 0 0 0 0
0 0 1 1 1 1 1 0 0
0 1 1 1 0 1 1 1 0
0 1 1 0 0 0 1 1 0
0 1 0 0 0 0 0 1 0
0 1 1 0 0 0 1 1 0
0 1 1 1 0 1 1 1 0
0 0 1 1 1 1 1 0 0
0 0 0 0 0 0 0 0 0
Then the indices of your sort of mask are, lets say [k n]
>> [k n]
ans =
3 2
4 2
5 2
6 2
7 2
2 3
3 3
4 3
6 3
7 3
8 3
2 4
3 4
7 4
8 4
2 5
8 5
2 6
3 6
7 6
8 6
2 7
3 7
4 7
6 7
7 7
8 7
3 8
4 8
5 8
6 8
7 8
Now have a 9x9 matrix of zeroes on hand called B, we can shift the whole thing over to the left by one pixel as follows using the formula (i+9*(j-1)) to convert double index to a single index.
>> B=zeros(9,9);
>> B((k)+9*(n-2))=1
B =
0 0 0 0 0 0 0 0 0
0 1 1 1 1 1 0 0 0
1 1 1 0 1 1 1 0 0
1 1 0 0 0 1 1 0 0
1 0 0 0 0 0 1 0 0
1 1 0 0 0 1 1 0 0
1 1 1 0 1 1 1 0 0
0 1 1 1 1 1 0 0 0
0 0 0 0 0 0 0 0 0
Or move down and to the right as follows
>> B=zeros(9,9);
>> B((k+1)+9*(n-0))=1
B =
0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0
0 0 0 1 1 1 1 1 0
0 0 1 1 1 0 1 1 1
0 0 1 1 0 0 0 1 1
0 0 1 0 0 0 0 0 1
0 0 1 1 0 0 0 1 1
0 0 1 1 1 0 1 1 1
0 0 0 1 1 1 1 1 0
As long as it doesn't go out of bounds you should be able to shift a single annular mask around with a simple addition to put the center at the image center.
I want to use a 3D blob detector which is a filtration algorithm, used for detection of spherical object in images. In this filter voxels with value greater than all pixels in their 26-neighborhood are set to 1
(s(x,y,z)=max(n26(x,y,z)))
and every other pixel is set to 0. Is there any function in matlab to do this work?
M(x,y,z)={ 1 if s(x,y,z)=max(n26(x,y,z))
0 otherwise
The easiest way to find local maxima is to use imdilate:
%# s = 3D array
msk = true(3,3,3);
msk(2,2,2) = false;
%# assign, to every voxel, the maximum of its neighbors
s_dil = imdilate(s,msk);
M = s > s_dil; %# M is 1 wherever a voxel's value is greater than its neighbors
matlabs own imregionalmax supports 26n from the get go, output is a logical.
2D example with 8n:
A =
1 1 1 1 1 1 1 1
1 3 3 3 1 1 4 1
1 3 5 3 1 4 4 4
1 3 3 3 1 4 4 4
1 1 1 1 1 4 6 4
1 1 1 1 1 4 4 4
>> B = imregionalmax(A);
>> B
B =
0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0
0 0 1 0 0 0 0 0
0 0 0 0 0 0 0 0
0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 0