Integers can only be combined with integers of the same class, or scalar doubles

Error using - Integers can only be combined with integers of the same class, or scalar doubles.

Error in Code (line 27) data = data - repmat(mn,1,N);

Here is the code snip, where I am subtracting off the mean of some matrices:

mn = mean(data,2);

data = data - repmat(mn,1,N);

I tried casting the new data matrix to an int16 but that gave the same error, not sure what to cast it to, to avoid error.

[jgg]

What is the datatype of the data object? This looks like your issue.

You might try data = double(data) first.

[jgg]

I added my answer below.

Are the matrices square? If they're not square, the * operator is matrix multiplication, if they're like NxM this is asking Matlab to do NxM * NxM which doesn't work. If you want to multiply them component-wise, the operator is .*.

See here for example.

[jgg]

I think you should post another question to cover this, and give more background. I don't really understand what you're trying to do here, so you might want to explain very precisely what you're trying to accomplish, and what the objects you're working with are. (For instance, why are you trying to get the eigenvalues of a non-square matrix? Which operation made it non-square; it wasn't subtracting off the means.)

Answer 1

[jgg]

Try setting data = double(data) first.

[Neo]

Oh wait, why did the matrix have to be casted into a double, I know because we didn't want the error, but why have an error in the first place?

[Chittaranjan Mohapatra]

Yes It Works

Answer 2

[sai kumar]

i have the same problem please help me!!.

Error using mod

Integers can only be combined with integers of the same class, or scalar doubles.

Error in trial>decryption_Callback (line 457) m1(l)=mod(c2,n); code: c1=reshape(c,1,[]); s=numel(c1); m1=c1; c2=c1; for l=1:s c2(l)=c1(l)^msk; m1(l)=mod(c2,n); end

Answer 3

[lokesh kumar]

Error???????????????????????

%rough dark channel computation%
%a=imread('D:\Image processing project\image\hazy.jpg');
a=imread('D:\Image processing project\contrast\dark channel prior\dehaze_code\dehaze_code\inputImgs\forest_input.png');

[ht,wt,~]=size(a);
patchSize=15;
padSize=7;
JDark=zeros(ht,wt);
imj=padarray(a,[padSize padSize],inf);
for i=1:ht
    for j=1:wt
        patch=imj(i:(i+padSize-1),j:(j+padSize-1),:);
        JDark(i,j)=min(patch(:));
    end
end

atmSum=zeros(1,3);
imsize=ht*wt;
numpx=floor(imsize/1000);
JDarkVec=reshape(JDark,imsize,1);
ImVec=reshape(a,imsize,3);

[JDarkVec,indices]=sort(JDarkVec);
indices=indices(imsize-numpx:end);

atmSum=zeros(1,3);
for ind=1:numpx
    atmSum=atmSum+ImVec(indices(ind),:);
end
A=atmSum/numpx;
figure;
imshow(uint8(A));

So what is the problem with the atmSum, it is not working. Anybody wants to solve this problem are always welcome.

or email me: [e-mail]

Thanks

[DGM]

Since the intermediate sum will exceed the dynamic range of the original integer class, do the accumulation in float. This avoids mixed class arithmetic problems, and it avoids truncation.

ImVec = double(ImVec);