Hi, I wanted to orthonormalize some transformation matricies which may become non-trivially non-ortho due to a discretization step. The interwebs suggest Gram-Schmidt orthonormalization but before I implement that I thought I would ask if there is a more efficient/stable/recommended way using already-coded Eigen capabilities.

Thanks

you can also use a QR factorization (Gram-Schmidt orthonormalization is one way to get a QR factorization):

Code: Select all

A_ortho = A.householderQr().householderQ();


I had a feeling you would have a rockstar answer

Thanks