An algorithm to compute the polar decomposition of a \(3 \times 3\) matrix (Q329306): Difference between revisions

The authors use an argument based on quaternions to explain how the orthogonal polar factor of a \(3 \times 3\) matrix can be obtained from an eigenvector corresponding to the dominant eigenvalue of a related \(4 \times 4\) symmetric matrix. This fact is then used to develop an algorithm for computing the polar decomposition of \(3 \times 3\) matrices. The authors show that the new algorithm requires fewer operations and is faster than the other methods based on the singular value decomposition.