I want to use a matrix to store grayscale values of multiple images. As such, my plan is to have a matrix with the dimension of axb where a = number of values in the vector where I temporarily store the grayscale values per image, b = number of images.

When I use MatrixXf mat(a,b), I get an Assertion failed error upon running the program. Can you anyone give me an insight as to how I may get around to declare it properly?

This is the error:

Code: Select all

pca: lib/eigen/Eigen/src/Core/Coeffs.h:96: typename Eigen::ei_traits<T>::Scalar& Eigen::MatrixBase<Derived>::operator()(int, int) [with Derived = Eigen::Matrix<int, 10000, 10000, 2, 10000, 10000>]: Assertion `row >= 0 && row < rows() && col >= 0 && col < cols()' failed.
Aborted


hm,

MatrixXf mat(a,b);

gives you a a x b matrix. The error you get means you tried to access mat(i,j) with i and/or j outside the ranges [0:a[, [0:b[. Without more information I cannot help you more.

Here is my code for accessing the matrix:

Code: Select all

void storeVector(){   
   int ctr=0;
   
   mat.resize(gScale.length(),mSize);

   for (int j=mSize; j<mSize+1; j++){   //cols
      for (int i=1; i<gScale.length(); i++){//rows
           mat(i,j) = gScale[ctr];
                ctr++;
        }
   }
    mSize++;
    gScale.empty(); // clear the arraylist
}



mSize is a global variable that keeps track of the number of rows. It is initially set at 1. The program still returns an Assert failure.

Indices go from (0,0) to (rows-1,columns-1)...

I realised my mistake. Thank you very much.

One other thing, how do I dynamically-resize the matrix without overwriting the contents of the previous matrix?

mat.conservativeResize(rows,cols);

Ok. Thanks. It's now working as it should.