JacobiSVD: implement general R-SVD using full-pivoting QR, so we now support any rectangular matrix size by reducing to the smaller of the two dimensions (which is also an optimization)

This commit is contained in:
Benoit Jacob 2009-09-02 06:36:55 -04:00
parent c16d65f015
commit e6b77bcc6b
4 changed files with 45 additions and 16 deletions

View File

@ -1,7 +1,7 @@
#ifndef EIGEN_SVD_MODULE_H
#define EIGEN_SVD_MODULE_H
#include "Core"
#include "QR"
#include "Householder"
#include "Jacobi"

View File

@ -286,7 +286,7 @@ FullPivotingHouseholderQR<MatrixType>& FullPivotingHouseholderQR<MatrixType>::co
m_cols_permutation.resize(matrix.cols());
int number_of_transpositions = 0;
RealScalar biggest;
RealScalar biggest(0);
for (int k = 0; k < size; ++k)
{

View File

@ -188,15 +188,42 @@ void ei_real_2x2_jacobi_svd(const MatrixType& matrix, int p, int q,
template<typename MatrixType, unsigned int Options>
JacobiSVD<MatrixType, Options>& JacobiSVD<MatrixType, Options>::compute(const MatrixType& matrix)
{
MatrixType work_matrix(matrix);
int size = matrix.rows();
if(ComputeU) m_matrixU = MatrixUType::Identity(size,size);
if(ComputeV) m_matrixV = MatrixUType::Identity(size,size);
m_singularValues.resize(size);
MatrixType work_matrix;
int rows = matrix.rows();
int cols = matrix.cols();
int diagSize = std::min(rows, cols);
if(ComputeU) m_matrixU = MatrixUType::Zero(rows,rows);
if(ComputeV) m_matrixV = MatrixVType::Zero(cols,cols);
m_singularValues.resize(diagSize);
const RealScalar precision = 2 * epsilon<Scalar>();
if(rows > cols)
{
FullPivotingHouseholderQR<MatrixType> qr(matrix);
work_matrix = qr.matrixQR().block(0,0,diagSize,diagSize).template triangularView<UpperTriangular>();
if(ComputeU) m_matrixU = qr.matrixQ();
if(ComputeV)
for(int i = 0; i < cols; i++)
m_matrixV.coeffRef(qr.colsPermutation().coeff(i),i) = Scalar(1);
}
else if(rows < cols)
{
FullPivotingHouseholderQR<MatrixType> qr(MatrixType(matrix.adjoint()));
work_matrix = qr.matrixQR().block(0,0,diagSize,diagSize).template triangularView<UpperTriangular>().adjoint();
if(ComputeV) m_matrixV = qr.matrixQ();
if(ComputeU)
for(int i = 0; i < rows; i++)
m_matrixU.coeffRef(qr.colsPermutation().coeff(i),i) = Scalar(1);
}
else
{
work_matrix = matrix;
if(ComputeU) m_matrixU.diagonal().setOnes();
if(ComputeV) m_matrixV.diagonal().setOnes();
}
sweep_again:
for(int p = 1; p < size; ++p)
for(int p = 1; p < diagSize; ++p)
{
for(int q = 0; q < p; ++q)
{
@ -219,27 +246,27 @@ sweep_again:
RealScalar biggestOnDiag = work_matrix.diagonal().cwise().abs().maxCoeff();
RealScalar maxAllowedOffDiag = biggestOnDiag * precision;
for(int p = 0; p < size; ++p)
for(int p = 0; p < diagSize; ++p)
{
for(int q = 0; q < p; ++q)
if(ei_abs(work_matrix.coeff(p,q)) > maxAllowedOffDiag)
goto sweep_again;
for(int q = p+1; q < size; ++q)
for(int q = p+1; q < diagSize; ++q)
if(ei_abs(work_matrix.coeff(p,q)) > maxAllowedOffDiag)
goto sweep_again;
}
for(int i = 0; i < size; ++i)
for(int i = 0; i < diagSize; ++i)
{
RealScalar a = ei_abs(work_matrix.coeff(i,i));
m_singularValues.coeffRef(i) = a;
if(ComputeU && (a!=RealScalar(0))) m_matrixU.col(i) *= work_matrix.coeff(i,i)/a;
}
for(int i = 0; i < size; i++)
for(int i = 0; i < diagSize; i++)
{
int pos;
m_singularValues.end(size-i).maxCoeff(&pos);
m_singularValues.end(diagSize-i).maxCoeff(&pos);
if(pos)
{
pos += i;

View File

@ -91,12 +91,14 @@ void test_jacobisvd()
CALL_SUBTEST( svd(Matrix3f()) );
CALL_SUBTEST( svd(Matrix4d()) );
CALL_SUBTEST( svd(MatrixXf(50,50)) );
// CALL_SUBTEST( svd(MatrixXd(14,7)) );
CALL_SUBTEST( svd(MatrixXcd(14,7)) );
CALL_SUBTEST( svd(MatrixXd(10,50)) );
CALL_SUBTEST( svd(MatrixXcf(3,3)) );
CALL_SUBTEST( svd(MatrixXd(30,30)) );
}
CALL_SUBTEST( svd(MatrixXf(200,200)) );
CALL_SUBTEST( svd(MatrixXcd(100,100)) );
CALL_SUBTEST( svd(MatrixXf(300,200)) );
CALL_SUBTEST( svd(MatrixXcd(100,150)) );
CALL_SUBTEST( svd_verify_assert<Matrix3f>() );
CALL_SUBTEST( svd_verify_assert<Matrix3d>() );