vpMatrix_svd.cpp

/****************************************************************************
 *
 * $Id: vpMatrix_svd.cpp 4210 2013-04-16 08:57:46Z fspindle $
 *
 * This file is part of the ViSP software.
 * Copyright (C) 2005 - 2013 by INRIA. All rights reserved.
 * 
 * This software is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * ("GPL") version 2 as published by the Free Software Foundation.
 * See the file LICENSE.txt at the root directory of this source
 * distribution for additional information about the GNU GPL.
 *
 * For using ViSP with software that can not be combined with the GNU
 * GPL, please contact INRIA about acquiring a ViSP Professional 
 * Edition License.
 *
 * See http://www.irisa.fr/lagadic/visp/visp.html for more information.
 * 
 * This software was developed at:
 * INRIA Rennes - Bretagne Atlantique
 * Campus Universitaire de Beaulieu
 * 35042 Rennes Cedex
 * France
 * http://www.irisa.fr/lagadic
 *
 * If you have questions regarding the use of this file, please contact
 * INRIA at visp@inria.fr
 * 
 * This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE
 * WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
 *
 *
 * Description:
 * Matrix SVD decomposition.
 *
 * Authors:
 * Eric Marchand
 *
 *****************************************************************************/

#include <visp/vpConfig.h>

#ifndef DOXYGEN_SHOULD_SKIP_THIS

#include <visp/vpMatrix.h>
#include <visp/vpMath.h>
#include <visp/vpColVector.h>
#include <visp/vpException.h>
#include <visp/vpMatrixException.h>
#include <visp/vpDebug.h>


#include <cmath>    // std::fabs
#include <limits>   // numeric_limits
#include <iostream>

/*---------------------------------------------------------------------

SVD related functions

---------------------------------------------------------------------*/


static double pythag(double a, double b)
{
  double absa, absb;
  absa = fabs(a);
  absb = fabs(b);
  if (absa > absb) return absa*sqrt(1.0+vpMath::sqr(absb/absa));
  //else return (absb == 0.0 ? 0.0 : absb*sqrt(1.0+vpMath::sqr(absa/absb)));
  else return (std::fabs(absb) <= std::numeric_limits<double>::epsilon() ? 0.0 : absb*sqrt(1.0+vpMath::sqr(absa/absb)));
}

#ifdef SIGN
#undef SIGN
#endif
#define SIGN(a,b) ((b) >= 0.0 ? fabs(a) : -fabs(a))

#define  MAX_ITER_SVD 50
void vpMatrix::svdNr(vpColVector& W, vpMatrix& V)
{

  unsigned int m = rowNum;
  unsigned int n = colNum;
  double epsilon = 10*std::numeric_limits<double>::epsilon();

  unsigned int flag,i,its,j,jj,k,l=0,nm=0;
  double c,f,h,s,x,y,z;
  double anorm=0.0,g=0.0,scale=0.0;

  // So that the original NRC code (using 1..n indexing) can be used
  // This should be considered as a temporary fix.
  double **a = new double*[m+1];
  double **v = new double*[n+1];
  //  double **w = W.rowPtrs;
  //  w--;

  double *w = new double[n+1];
  for (i=0;i<n;i++) w[i+1] = 0.0;

  for (i=1;i<=m;i++) {
    a[i] = this->rowPtrs[i-1]-1;
  }
  for (i=1;i<=n;i++) {
    v[i] = V.rowPtrs[i-1]-1;
  }

  if (m < n)
  {
    delete[] w;
    delete[] a;
    delete[] v;
    vpERROR_TRACE("\n\t\tSVDcmp: You must augment A with extra zero rows") ;
    throw(vpMatrixException(vpMatrixException::matrixError,
                "\n\t\tSVDcmp: You must augment A with "
                "extra zero rows")) ;
  }
  double* rv1=new double[n+1];

  for (i=1;i<=n;i++) {
    l=i+1;
    rv1[i]=scale*g;
    g=s=scale=0.0;
    if (i <= m) {
      for (k=i;k<=m;k++) scale += fabs(a[k][i]);
      //if ((scale != 0.0) || (fabs(scale) > EPS_SVD)) {
      if ((std::fabs(scale) > epsilon)/* || (fabs(scale) > EPS_SVD)*/) {
    for (k=i;k<=m;k++) {
      a[k][i] /= scale;
      s += a[k][i]*a[k][i];
    }
    f=a[i][i];
    g = -SIGN(sqrt(s),f);
    h=f*g-s;
    a[i][i]=f-g;
    if (i != n) {
      for (j=l;j<=n;j++) {
        for (s=0.0,k=i;k<=m;k++) s += a[k][i]*a[k][j];
        f=s/h;
        for (k=i;k<=m;k++) a[k][j] += f*a[k][i];
      }
    }
    for (k=i;k<=m;k++) a[k][i] *= scale;
      }
    }
    w[i]=scale*g;
    g=s=scale=0.0;
    if (i <= m && i != n) {
      for (k=l;k<=n;k++) scale += fabs(a[i][k]);
      //if ((scale != 0.0) || (fabs(scale) > EPS_SVD)) {
      if ((std::fabs(scale) > epsilon) /*|| (fabs(scale) > EPS_SVD)*/) {
    for (k=l;k<=n;k++) {
      a[i][k] /= scale;
      s += a[i][k]*a[i][k];
    }
    f=a[i][l];
    g = -SIGN(sqrt(s),f);
    h=f*g-s;
    a[i][l]=f-g;
    for (k=l;k<=n;k++) rv1[k]=a[i][k]/h;
    if (i != m) {
      for (j=l;j<=m;j++) {
        for (s=0.0,k=l;k<=n;k++) s += a[j][k]*a[i][k];
        for (k=l;k<=n;k++) a[j][k] += s*rv1[k];
      }
    }
    for (k=l;k<=n;k++) a[i][k] *= scale;
      }
    }
    anorm=vpMath::maximum(anorm,(fabs(w[i])+fabs(rv1[i])));
  }
  for (i=n;i>=1;i--) {
    if (i < n) {
      //if ((g) || (fabs(g) > EPS_SVD)) {
      if ((std::fabs(g) > epsilon) /*|| (fabs(g) > EPS_SVD)*/) {
    for (j=l;j<=n;j++)
      v[j][i]=(a[i][j]/a[i][l])/g;
    for (j=l;j<=n;j++) {
      for (s=0.0,k=l;k<=n;k++) s += a[i][k]*v[k][j];
      for (k=l;k<=n;k++) v[k][j] += s*v[k][i];
    }
      }
      for (j=l;j<=n;j++) v[i][j]=v[j][i]=0.0;
    }
    v[i][i]=1.0;
    g=rv1[i];
    l=i;
  }
  for (i=n;i>=1;i--) {
    l=i+1;
    g=w[i];
    if (i < n)
      for (j=l;j<=n;j++) a[i][j]=0.0;
    //if ((g != 0.0) || (fabs(g) > EPS_SVD)) {
    if ((std::fabs(g) > epsilon) /*|| (fabs(g) > EPS_SVD)*/) {
      g=1.0/g;
      if (i != n) {
    for (j=l;j<=n;j++) {
      for (s=0.0,k=l;k<=m;k++) s += a[k][i]*a[k][j];
      f=(s/a[i][i])*g;
      for (k=i;k<=m;k++) a[k][j] += f*a[k][i];
    }
      }
      for (j=i;j<=m;j++) a[j][i] *= g;
    } else {
      for (j=i;j<=m;j++) a[j][i]=0.0;
    }
    ++a[i][i];
  }
  for (k=n;k>=1;k--) {
    for (its=1;its<=MAX_ITER_SVD;its++) {
      flag=1;
      for (l=k;l>=1;l--) {
    nm=l-1;
    //if ((fabs(rv1[l])+anorm == anorm) || (fabs(rv1[l]) <= EPS_SVD)) {
        if ((std::fabs(rv1[l]) <= epsilon) /*|| (fabs(rv1[l]) <= EPS_SVD)*/) {
      flag=0;
      break;
    }
    //if ((fabs(w[nm])+anorm == anorm) || (fabs(w[nm]) <= EPS_SVD)) break;
        if ((std::fabs(w[nm]) <= epsilon) /*|| (fabs(w[nm]) <= EPS_SVD)*/) break;
      }
      if (flag) {
    c=0.0;
    s=1.0;
    for (i=l;i<=k;i++) {
      f=s*rv1[i];
      //if ((fabs(f)+anorm != anorm)  || (fabs(f) <= EPS_SVD)) {
          if ((std::fabs(f) > epsilon)  /*|| (fabs(f) <= EPS_SVD)*/) {
        g=w[i];
        h=pythag(f,g);
        w[i]=h;
        h=1.0/h;
        c=g*h;
        s=(-f*h);
        for (j=1;j<=m;j++) {
          y=a[j][nm];
          z=a[j][i];
          a[j][nm]=y*c+z*s;
          a[j][i]=z*c-y*s;
        }
      }
    }
      }
      z=w[k];
      if (l == k) {
    if (z < 0.0) {
      w[k] = -z;
      for (j=1;j<=n;j++) v[j][k]=(-v[j][k]);
    }
    break;
      }
      if (its == MAX_ITER_SVD)
      {
    for (i=0;i<n;i++) W[i] = w[i+1];

    vpERROR_TRACE("\n\t\t No convergence in  SVDcmp ") ;
    std::cout << *this <<std::endl ;
    //  throw(vpMatrixException(vpMatrixException::matrixError,
    //              "\n\t\t No convergence in  SVDcmp ")) ;
      }
      x=w[l];
      nm=k-1;
      y=w[nm];
      g=rv1[nm];
      h=rv1[k];
      f=((y-z)*(y+z)+(g-h)*(g+h))/(2.0*h*y);
      g=pythag(f,1.0);
      f=((x-z)*(x+z)+h*((y/(f+SIGN(g,f)))-h))/x;
      c=s=1.0;
      for (j=l;j<=nm;j++) {
    i=j+1;
    g=rv1[i];
    y=w[i];
    h=s*g;
    g=c*g;
    z=pythag(f,h);
    rv1[j]=z;
    c=f/z;
    s=h/z;
    f=x*c+g*s;
    g=g*c-x*s;
    h=y*s;
    y=y*c;
    for (jj=1;jj<=n;jj++) {
      x=v[jj][j];
      z=v[jj][i];
      v[jj][j]=x*c+z*s;
      v[jj][i]=z*c-x*s;
    }
    z=pythag(f,h);
    w[j]=z;
    //if ((z != 0.0) || (fabs(z) > EPS_SVD)) {
        if ((std::fabs(z) > epsilon) /*|| (fabs(z) > EPS_SVD)*/) {
      z=1.0/z;
      c=f*z;
      s=h*z;
    }
    f=(c*g)+(s*y);
    x=(c*y)-(s*g);
    for (jj=1;jj<=m;jj++) {
      y=a[jj][j];
      z=a[jj][i];
      a[jj][j]=y*c+z*s;
      a[jj][i]=z*c-y*s;
    }
      }
      rv1[l]=0.0;
      rv1[k]=f;
      w[k]=x;
    }
  }
  for (i=0;i<n;i++) W[i] = w[i+1];


  delete[] w;
  delete[] rv1;
  delete[] a;
  delete[] v;

}

#undef SIGN
#undef PYTHAG

void vpMatrix::SVBksb( const vpColVector& w,
               const vpMatrix& v,
               const vpColVector& b, vpColVector& x)
{
  unsigned int m = this->rowNum;
  unsigned int n = this->colNum;
  double** u = rowPtrs;

  unsigned int jj,j,i;
  double s,*tmp;

  tmp=new double[n];
  for (j=0;j<n;j++) {
    s=0.0;
    //if (w[j])
    if (std::fabs(w[j]) > std::numeric_limits<double>::epsilon())
    {
      for (i=0;i<m;i++) s += u[i][j]*b[i];
      s /= w[j];
    }
    tmp[j]=s;
  }
  for (j=0;j<n;j++) {
    s=0.0;
    for (jj=0;jj<n;jj++) s += v[j][jj]*tmp[jj];
    x[j]=s;
  }
  delete [] tmp;
}

#define TOL 1.0e-5

#define TOLERANCE 1.0e-7

static
void svd_internal_use(double *U, double *S, double *V,
              unsigned int nRow, unsigned int nCol)
{
  unsigned int i, j, k, EstColRank, RotCount, SweepCount, slimit;
  double eps, e2, tol, vt, p, x0, y0, q, r, c0, s0, d1, d2;

  eps = TOLERANCE;
  slimit = nCol / 4;
  if (slimit < 6.0)
    slimit = 6;
  SweepCount = 0;
  e2 = 10.0 * nRow * eps * eps;
  tol = eps * .1;
  EstColRank = nCol;
  if(V)
    for (i = 0; i < nCol; i++)
      for (j = 0; j < nCol; j++) {
    V[nCol * i + j] = 0.0;
    V[nCol * i + i] = 1.0;
      }
  RotCount = EstColRank * (EstColRank - 1) / 2;
  while (RotCount != 0 && SweepCount <= slimit) {
    RotCount = EstColRank * (EstColRank - 1) / 2;
    SweepCount++;
    for (j = 0; j < EstColRank - 1; j++) {
      for (k = j + 1; k < EstColRank; k++) {
    p = q = r = 0.0;
    for (i = 0; i < nRow; i++) {
      x0 = U[nCol * i + j];
      y0 = U[nCol * i + k];
      p += x0 * y0;
      q += x0 * x0;
      r += y0 * y0;
    }
    S[j] = q;
    S[k] = r;
    if (q >= r) {
      if (q <= e2 * S[0] || fabs(p) <= tol * q)
        RotCount--;
      else {
        p /= q;
        r = 1 - r / q;
        vt = sqrt(4 * p * p + r * r);
        c0 = sqrt(fabs(.5 * (1 + r / vt)));
        s0 = p / (vt * c0);
        for (i = 0; i < nRow; i++) {
          d1 = U[nCol * i + j];
          d2 = U[nCol * i + k];
          U[nCol * i + j] = d1 * c0 + d2 * s0;
          U[nCol * i + k] = -d1 * s0 + d2 * c0;
        }
        if(V)
          for (i = 0; i < nCol; i++) {
        d1 = V[nCol * i + j];
        d2 = V[nCol * i + k];
        V[nCol * i + j] = d1 * c0 + d2 * s0;
        V[nCol * i + k] = -d1 * s0 + d2 * c0;
          }
      }
    }
    else {
      p /= r;
      q = q / r - 1;
      vt = sqrt(4 * p * p + q * q);
      s0 = sqrt(fabs(.5 * (1 - q / vt)));
      if (p < 0)
        s0 = -s0;
      c0 = p / (vt * s0);
      for (i = 0; i < nRow; i++) {
        d1 = U[nCol * i + j];
        d2 = U[nCol * i + k];
        U[nCol * i + j] = d1 * c0 + d2 * s0;
        U[nCol * i + k] = -d1 * s0 + d2 * c0;
      }
      if(V)
        for (i = 0; i < nCol; i++) {
          d1 = V[nCol * i + j];
          d2 = V[nCol * i + k];
          V[nCol * i + j] = d1 * c0 + d2 * s0;
          V[nCol * i + k] = -d1 * s0 + d2 * c0;
        }
    }
      }
    }
    while (EstColRank >= 3 && S[(EstColRank - 1)] <= S[0] * tol + tol * tol)
      EstColRank--;
  }
  for(i = 0; i < nCol; i++)
    S[i] = sqrt(S[i]);
  for(i = 0; i < nCol; i++)
    for(j = 0; j < nRow; j++)
      U[nCol * j + i] = U[nCol * j + i] / S[i];
}

void vpMatrix::svdFlake(vpColVector &W, vpMatrix &V)
{


  svd_internal_use(data, W.data, V.data, getRows(), getCols());
}


#if (VISP_HAVE_OPENCV_VERSION >= 0x020101) // Require opencv >= 2.1.1
#    include <opencv2/core/core.hpp>
void vpMatrix::svdOpenCV(vpColVector& w, vpMatrix& v){
  int rows = (int)this->getRows();
  int cols = (int)this->getCols();
  cv::Mat m(rows, cols, CV_64F, this->data);
  cv::SVD opencvSVD(m);
  cv::Mat opencvV = opencvSVD.vt;
  cv::Mat opencvW = opencvSVD.w;
  v.resize((unsigned int)opencvV.rows, (unsigned int)opencvV.cols);
  w.resize((unsigned int)(opencvW.rows*opencvW.cols));
  
  memcpy(v.data, opencvV.data, (size_t)(8*opencvV.rows*opencvV.cols));
  v=v.transpose();
  memcpy(w.data, opencvW.data, (size_t)(8*opencvW.rows*opencvW.cols));
  this->resize((unsigned int)opencvSVD.u.rows, (unsigned int)opencvSVD.u.cols);
  memcpy(this->data,opencvSVD.u.data, (size_t)(8*opencvSVD.u.rows*opencvSVD.u.cols));
}

#endif

#ifdef VISP_HAVE_LAPACK
extern "C" int dgesdd_(char *jobz, int *m, int *n, double *a, int *lda, double *s, double *u, int *ldu, double *vt, int *ldvt, double *work, int *lwork, int *iwork, int *info);
#include <stdio.h>
#include <string.h>

void vpMatrix::svdLapack(vpColVector& W, vpMatrix& V){
  /* unsigned */ int m = static_cast<int>(this->getCols()), n = static_cast<int>(this->getRows()), lda = m, ldu = m, ldvt = std::min(m,n);
  int info, lwork;

  double wkopt;
  double* work;

  int* iwork = new int[8*static_cast<unsigned int>(std::min(n,m))];

  double *s = W.data;
  double* a = new double[static_cast<unsigned int>(lda*n)];
  memcpy(a,this->data,this->getRows()*this->getCols()*sizeof(double));
  double* u = V.data;
  double* vt = this->data;



  lwork = -1;
  dgesdd_( (char*)"S", &m, &n, a, &lda, s, u, &ldu, vt, &ldvt, &wkopt, &lwork, iwork, &info );
  lwork = (int)wkopt;
  work = new double[static_cast<unsigned int>(lwork)];

  dgesdd_( (char*)"S", &m, &n, a, &lda, s, u, &ldu, vt, &ldvt, work, &lwork, iwork, &info );

  if( info > 0 ) {
    std::cout << "The algorithm computing SVD failed to converge." << std::endl;

  }

  V=V.transpose();
  delete[] work;
  delete[] iwork;
  delete[] a;
}
#endif

#ifdef VISP_HAVE_GSL
#include <gsl/gsl_linalg.h>

void
vpMatrix::svdGsl(vpColVector& w, vpMatrix& v)
{
  
#if 0 
  // premier test avec la gsl 1. on recopie...
  int i,j ;

  int nc = getCols() ;
  int nr = getRows() ;
  gsl_matrix *A = gsl_matrix_alloc(nr, nc) ;

  int Atda = A->tda ;
  for (i=0 ; i < nr ; i++)
  {
    int k = i*Atda ;
    for (j=0 ; j < nc ; j++)
      A->data[k+j] = (*this)[i][j] ;
  }
  // gsl_matrix_set(A,i,j,(*this)[i][j]) ;

  gsl_matrix *V = gsl_matrix_alloc(nc, nc) ;
  gsl_vector *S = gsl_vector_alloc(nc) ;
  gsl_vector *work = gsl_vector_alloc(nc) ;

  gsl_linalg_SV_decomp(A,V,S, work) ;
//  gsl_linalg_SV_decomp_jacobi(A,V,S) ;


  //l'acces par gsl_matrix_get est tres lourd, voir si on peut pas faire
  // autremement (surement !)

  Atda = A->tda ;
  for (i=0 ; i < nr ; i++)
    for (j=0 ; j < nc ; j++)
      (*this)[i][j] =  gsl_matrix_get(A,i,j) ;

  int Vtda = V->tda ;
  for (i=0 ; i < nc ; i++)
  {
    int k = i*Vtda ;
    for (j=0 ; j < nc ; j++)
      v[i][j] = V->data[k+j] ;
  }

  for (j=0 ; j < nc ; j++)
    w[j] = gsl_vector_get(S,j) ;


  gsl_matrix_free(V) ;
  gsl_matrix_free(A) ;
  gsl_vector_free(S) ;
  gsl_vector_free(work) ;

#else //optimisation Anthony 20/03/2008
  
  unsigned int nc = getCols() ;
  unsigned int nr = getRows() ;
  gsl_vector *work = gsl_vector_alloc(nc) ;

//  gsl_linalg_SV_decomp_jacobi(A,V,S) ;


  //l'acces par gsl_matrix_get est tres lourd, voir si on peut pas faire
  // autremement (surement !)

  gsl_matrix A;
  A.size1 = nr;
  A.size2 = nc;
  A.tda = A.size2;
  A.data = this->data;
  A.owner = 0;
  A.block = 0;
  
  gsl_matrix V;
  V.size1 = nc;
  V.size2 = nc;
  V.tda = V.size2;
  V.data = v.data;
  V.owner = 0;
  V.block = 0;
  
  gsl_vector S;
  S.size = nc;
  S.stride = 1;
  S.data = w.data;
  S.owner = 0;
  S.block = 0;
  
  gsl_linalg_SV_decomp(&A,&V,&S, work) ;
  
  gsl_vector_free(work) ;

#endif  
}
#endif // # #GSL


#undef TOL
#undef TOLERANCE

#undef MAX_ITER_SVD

#endif // doxygen should skip this