How to use calcCovarMatrix on IplImage? - visual-c++

Question

There's a single-channel grayscale IplImage whose covariance matrix is to be calculated. There does happen to be a similar question on SO but no one has answered it and the code is significantly different.
Here's the code that's throwing an "unhandled exception":
int calcCovar( IplImage *src, float* dst, int w )
{
// Input matrix size
int rows = w;
int cols = w;
int i,j;
CvScalar se;
float *a;
a = (float*)malloc( w * w * sizeof(float) );
long int k=0;
//image pixels into 1D array
for(i = 0; i < w; ++i)
{
for(j = 0; j < w; ++j)
{
se = cvGet2D(src, i, j);
a[k++] = (float)se.val[0];
}
}
CvMat input = cvMat(w,w,CV_32FC1, a); //Is this the right way to format input pixels??
// Covariance matrix is N x N,
// where N is input matrix column size
const int n = w;
// Output variables passed by reference
CvMat* output = cvCreateMat(n, n, CV_32FC1);
CvMat* meanvec = cvCreateMat(1, rows, CV_32FC1);
// Calculate covariance matrix - error is here!!
cvCalcCovarMatrix((const void **) &input, rows, output, meanvec, CV_COVAR_NORMAL);
k = 0;
//Show result
cout << "Covariance matrix:" << endl;
for(i=0; i<n; i++) {
for(j=0; j<n; j++) {
cout << "(" << i << "," << j << "): ";
printf ("%f ", cvGetReal2D(output,i,j) / (rows - 1));
dst[k++] = cvGetReal2D(output,i,j) / (rows - 1);
//cout << "\t";
}
cout << endl;
}
return(0);
}

Answer 1

read in the manual about this function. if you store your values in the single matrix 'input', the second function parameter must not be 'rows'. The second parameter is used to say how many matrices you passes in the first parameter. You have passed just one matrix, 'input'. Segfault is no surprise.