please help me with this MATLAB error: To RESHAPE the number of elements must not change.

clc;
close all;
clear all;
x = imread('frame1.jpg');
x=x(:,:,1);
y = medfilt2(x); % required to find the correlation of the same image 
xmean=mean(x(:));
ymean=mean(y(:));

xnew=x-xmean; % finding mean for correlation matrix
ynew=y-ymean;
cm=normxcorr2(xnew,ynew);  % correlation matrix
s=size(cm);
n=min(s);
cm_square=cm(1:n,1:n);
[V,D] = eig(cm_square);  % Finding Eigenvectors

   cm_eig_vec = []; % Sorting Eigenvectors 
   eigValue=diag(D);
   [eigValue,IX]=sort(eigValue,'descend');
   cm_eig_vec=V(:,IX);

   norm_eigVector=sqrt(sum(cm_eig_vec.^2)); % normailization 
   cm_eig_vec=cm_eig_vec./repmat(norm_eigVector,size(cm_eig_vec,1),1);

   Eigenfaces = xmean * cm_eig_vec(:,1:1); % 1:dimensions (dimensionality reduction)
   Ipc1 = reshape(Eigenfaces(:,1),size(x,1),size(x,2));

Answer 1

[Walter Roberson]

Let L be the smaller of the number of rows and columns of your image -- as in

YourImage = imread('frame1.jpg');
[r, c, p] = size(YourImage);
L = min([r, c]);

I used different variable names here because you write over your x variable and I did not want any ambiguity over which size was being discussed.

Then, your cm_eig_vec variable is going to come out as (2*L-1) x (2*L-1), and your Eigenfaces variable is going to come out as (2*L-1) x 1.

You then try to reshape that vector of length (2*L-1) into the full rows and columns of the image, r x c. This cannot work unless your original image was 1 x 1.

[Harpreet Kaur]

you are right, it works for 1x1. But I don't need a image with 1x1 dimensions. I have to process it further for detecting forgery.

[Walter Roberson]

Well, by the time you get to cm_eig_vec you have a variable which is roughly twice as many rows and twice as many columns as the minimum dimension of your original matrix, but you then take only a single column of that to construct Eigenfaces. There is no way you can get the larger dimension back by just reshape() of that result.