#include <itkMathSVD.h>
Result of a runtime-sized SVD (square or rectangular): A == U diag(W) V^T, W descending. For an m x n input, U is m x k, V is n x k and W has length k = min(m, n) (thin factors). rcond < 0 auto-selects a k*epsilon threshold.
Definition at line 165 of file itkMathSVD.h.
Public Member Functions | |
| vnl_matrix< TReal > | PseudoInverse (TReal rcond=TReal{ -1 }) const |
| unsigned int | Rank (TReal rcond=TReal{ -1 }) const |
| vnl_matrix< TReal > | Recompose (TReal rcond=TReal{ -1 }) const |
| vnl_vector< TReal > | Solve (const vnl_vector< TReal > &b, TReal rcond=TReal{ -1 }) const |
Public Attributes | |
| vnl_matrix< TReal > | U {} |
| vnl_matrix< TReal > | V {} |
| vnl_vector< TReal > | W {} |
|
inline |
Moore-Penrose pseudo-inverse A^+.
Definition at line 173 of file itkMathSVD.h.
|
inline |
Numerical rank (count of singular values above rcond*max(w)).
Definition at line 187 of file itkMathSVD.h.
|
inline |
Reconstruct A with singular values at or below rcond*max(w) zeroed.
Definition at line 194 of file itkMathSVD.h.
|
inline |
Least-squares / minimum-norm solution of A x = b.
Definition at line 180 of file itkMathSVD.h.
| vnl_matrix<TReal> itk::Math::SVDResult< TReal >::U {} |
Definition at line 167 of file itkMathSVD.h.
Referenced by itk::Math::SVD().
| vnl_matrix<TReal> itk::Math::SVDResult< TReal >::V {} |
Definition at line 169 of file itkMathSVD.h.
Referenced by itk::Math::SVD().
| vnl_vector<TReal> itk::Math::SVDResult< TReal >::W {} |
Definition at line 168 of file itkMathSVD.h.
Referenced by itk::Math::SVD().