UCBsplines.h

//===========================================================================
// SINTEF Multilevel B-spline Approximation library - version 1.1
//
// Copyright (C) 2000-2005 SINTEF ICT, Applied Mathematics, Norway.
//
// This program is free software; you can redistribute it and/or          
// modify it under the terms of the GNU General Public License            
// as published by the Free Software Foundation version 2 of the License. 
//
// This program is distributed in the hope that it will be useful,        
// but WITHOUT ANY WARRANTY; without even the implied warranty of         
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the          
// GNU General Public License for more details.                           
//
// You should have received a copy of the GNU General Public License      
// along with this program; if not, write to the Free Software            
// Foundation, Inc.,                                                      
// 59 Temple Place - Suite 330,                                           
// Boston, MA  02111-1307, USA.                                           
//
// Contact information: e-mail: tor.dokken@sintef.no                      
// SINTEF ICT, Department of Applied Mathematics,                         
// P.O. Box 124 Blindern,                                                 
// 0314 Oslo, Norway.                                                     
//
// Other licenses are also available for this software, notably licenses
// for:
// - Building commercial software.                                        
// - Building software whose source code you wish to keep private.        
//===========================================================================
#ifndef _UNIF_QUBIC_BSPLINES_H_
#define _UNIF_QUBIC_BSPLINES_H_

#ifdef WIN32 
#define WIN32ORSGI6 
#endif

#ifdef SGI6  // roxar definition
#define WIN32ORSGI6
#endif

#ifdef SGI
#define WIN32ORSGI6
#endif

#include <GenMatrix.h>
#include <UCBtypedef.h>

//#include <cmath>
#include <math.h>

namespace UCBspl {

// Controls compilation for cubic C1 splines
//old: #define MBA_CUBIC_C1 1
  
//#define UNIFORM_CUBIC_C1_SPLINES 1

  
#ifdef  UNIFORM_CUBIC_C1_SPLINES
  // Basis functions
  inline double B_0(double t) {double tmp = 1.0-t; return tmp*tmp*tmp/2.0;}
  inline double B_1(double t) {return (t*(t*(2.5*t - 4.5) + 1.5) + 0.5);}
  inline double B_2(double t) {return (t*t*(-2.5*t + 3.0));}
  inline double B_3(double t) {return t*t*t/2.0;}
  
  // And their derivatives
  inline double dB_0(double t) {return 1.5*(1.0-t)*(t-1.0);}
  inline double dB_1(double t) {return (t*(7.5*t - 9.0) + 1.5);}
  inline double dB_2(double t) {return (t*(-7.5*t + 6.0));}
  inline double dB_3(double t) {return 1.5*t*t;}
  
  // PRE EVALUATTED tensor products of B-splines, in a grid index
  // -------------------------------------------------------------
  
  // Pre-evaluated tensors equivalent to B(k,s)*B(l,t)
  static double tensor_BB[2][2] = {
    {1./4., 1./4.},
    {1./4., 1./4.}
  };
  
  // ???? only temp. for compilation
  // equivalent to dB(k,s)*B(l,t).
  static double tensor_dBB[2][2] = {
    {-3./4., -3./4.},
    { 3./4.,  3./4.}
  };
  
  // The transpose of the matrix above, for the y-components,
  // equivalent to B(k,s)*dB(l,t)
  static double tensor_BdB[2][2] = {
    {-3./4., 3./4.},
    {-3./4., 3./4.}
  };
  
#else
  // The uniform (cardinal), univariate cubic C2 basis functions
  // 0 <= t < 1;
  inline double B_0(double t) {double tmp = 1.0-t; return tmp*tmp*tmp/6.0;}
  
  //static inline double B_1(double t) {return (3.0*t*t*t - 6.0*t*t+4.0)/6.0;}
  static double div46 = 4.0/6.0;
  inline double B_1(double t) {return (0.5*t*t*t - t*t + div46);}
  
  //static inline double B_2(double t) {return (-3.0*t*t*t + 3.0*t*t + 3.0*t + 1.0)/6.0;}
  static double div16 = 1.0/6.0;
  inline double B_2(double t) {return (-0.5*t*t*t + 0.5*t*t + 0.5*t + div16);}
  
  inline double B_3(double t) {return t*t*t/6.0;}
  
  // FIRST DERIVATIVES to the uniform cubic C2 basis functions
  inline double dB_0(double t) {return 0.5*(1.0-t)*(t-1.0);}
  inline double dB_1(double t) {return (1.5*t*t - 2.0*t);}
  inline double dB_2(double t) {return (-(1.5*t*t) + t + 0.5);}
  inline double dB_3(double t) {return t*t/2.0;}
  
  // Lagt til, Vegard (But not for C1 yet)
  // SECOND DERIVATIVES to the uniform cubic C2 basis functions
  inline double ddB_0(double t) {return (-t+1.0);}
  inline double ddB_1(double t) {return (3.0*t-2.0);}
  inline double ddB_2(double t) {return (-3.0*t+1.0);}
  inline double ddB_3(double t) {return t;}
  
  // PRE EVALUATTED tensor products of B-splines, in a grid index
  // -------------------------------------------------------------
  
  // Pre-evaluated tensors equivalent to B(k,s)*B(l,t)
  static double tensor_BB[3][3] = {
    {1./36., 1./9., 1./36.},
    {1./9., 4./9., 1./9.},
    {1./36., 1./9., 1./36.}
  };
  
  
  // Pre-evaluated tensor products of B-splines, in a grid index
  // Mixed derivative an positional tensor for the x-component.
  // equivalent to dB(k,s)*B(l,t).
  // (The second row with zeros is not used in evaluations.)
  static double tensor_dBB[3][3] = {
    {-1./12., -1./3., -1./12.},
    {     0.,     0.,      0.},
    { 1./12.,  1./3.,  1./12.}
  };
  
  // The transpose of the matrix above, for the y-components,
  // equivalent to B(k,s)*dB(l,t)
  static double tensor_BdB[3][3] = {
    {-1./12., 0., 1./12.},
    { -1./3., 0., 1./3.},
    {-1./12., 0., 1./12.}
  };
  
#endif
  
  
  // Generic interface for evaluation to all univariate basis functions
  inline double B(int k, double t) {
    if (k == 0)
      return B_0(t);
    else if (k == 1)
      return B_1(t);
    else if (k == 2)
      return B_2(t);
    else
      return B_3(t);
  }
  
  // Generic interface to first derivatives
  inline double dB(int k, double t) {
    if (k == 0)
      return dB_0(t);
    else if (k == 1)
      return dB_1(t);
    else if (k == 2)
      return dB_2(t);
    else
      return dB_3(t);
  }
  
  // Lagt til, Vegard
  // Generic interface to second derivatives
  
  inline double ddB(int k, double t) {
    if (k == 0)
      return ddB_0(t);
    else if (k == 1)
      return ddB_1(t);
    else if (k == 2)
      return ddB_2(t);
    else
      return ddB_3(t);
  }
  
  // Evaluation of a bivariate tensor product basis function
  inline double w(int k, int l, double s, double t) {
    return B(k,s)*B(l,t);
  }
  
  // s,t \in [0,1) (but special on gridlines m and n)
  // i,j \in [-1, ???
  inline void ijst(int m, int n, double uc, double vc, int& i, int& j, double& s, double& t) {
    //int i = std::min((int)uc - 1, m-2);
    //int j = std::min((int)vc - 1, n-2);
    
#ifdef  UNIFORM_CUBIC_C1_SPLINES
    i = 2*((int)uc) - 1;
    j = 2*((int)vc) - 1;
#else
    i = (int)uc - 1;
    j = (int)vc - 1;
#endif
    
    s = uc - floor(uc);
    t = vc - floor(vc);
    
    // adjust for x or y on gridlines m and n (since impl. has 0 <= x <= m and 0 <= y <= n 
#ifdef  UNIFORM_CUBIC_C1_SPLINES
    if (i == 2*m-1) {
      i-=2;
      s = 1;
    }
    if (j == 2*n-1) {
      j-=2;
      t = 1;
    }
#else
    if (i == m-1) {
      i--;
      s = 1;
    }
    if (j == n-1) {
      j--;
      t = 1;
    }
#endif
  }

inline void WKL(double s, double t, double w_kl[4][4])
{
    double Bs0 = B_0(s); double Bt0 = B_0(t);
    double Bs1 = B_1(s); double Bt1 = B_1(t);
    double Bs2 = B_2(s); double Bt2 = B_2(t);
    double Bs3 = B_3(s); double Bt3 = B_3(t);

    w_kl[0][0] = Bs0 * Bt0;
    w_kl[0][1] = Bs0 * Bt1;
    w_kl[0][2] = Bs0 * Bt2;
    w_kl[0][3] = Bs0 * Bt3;     

    w_kl[1][0] = Bs1 * Bt0;      
    w_kl[1][1] = Bs1 * Bt1;     
    w_kl[1][2] = Bs1 * Bt2;     
    w_kl[1][3] = Bs1 * Bt3;      

    w_kl[2][0] = Bs2 * Bt0;     
    w_kl[2][1] = Bs2 * Bt1;
    w_kl[2][2] = Bs2 * Bt2;     
    w_kl[2][3] = Bs2 * Bt3;     

    w_kl[3][0] = Bs3 * Bt0;     
    w_kl[3][1] = Bs3 * Bt1;     
    w_kl[3][2] = Bs3 * Bt2;     
    w_kl[3][3] = Bs3 * Bt3;     

}
  
inline void WKLandSum2(double s, double t, double w_kl[4][4], double& sum_w_ab2) 
{
    sum_w_ab2 = 0.0;
    double Bs0 = B_0(s); double Bt0 = B_0(t);
    double Bs1 = B_1(s); double Bt1 = B_1(t);
    double Bs2 = B_2(s); double Bt2 = B_2(t);
    double Bs3 = B_3(s); double Bt3 = B_3(t);

    double tmp;

    // unrolled by Odd Andersen 15. dec. 2003, for optimization
    tmp = Bs0 * Bt0; w_kl[0][0] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs0 * Bt1; w_kl[0][1] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs0 * Bt2; w_kl[0][2] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs0 * Bt3; w_kl[0][3] = tmp; sum_w_ab2 += tmp * tmp;

    tmp = Bs1 * Bt0; w_kl[1][0] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs1 * Bt1; w_kl[1][1] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs1 * Bt2; w_kl[1][2] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs1 * Bt3; w_kl[1][3] = tmp; sum_w_ab2 += tmp * tmp;

    tmp = Bs2 * Bt0; w_kl[2][0] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs2 * Bt1; w_kl[2][1] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs2 * Bt2; w_kl[2][2] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs2 * Bt3; w_kl[2][3] = tmp; sum_w_ab2 += tmp * tmp;

    tmp = Bs3 * Bt0; w_kl[3][0] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs3 * Bt1; w_kl[3][1] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs3 * Bt2; w_kl[3][2] = tmp; sum_w_ab2 += tmp * tmp;
    tmp = Bs3 * Bt3; w_kl[3][3] = tmp; sum_w_ab2 += tmp * tmp;

//     int k,l;
//     sum_w_ab2 = 0.0; 
//     for (k = 0; k <= 3; k++) {
//      for (l = 0; l <=3; l++) {
//          double tmp = w(k, l, s, t);
//          w_kl[k][l] = tmp;
//          sum_w_ab2 += (tmp*tmp); 
//      }
//     }
}
    
  // for check, should give unity
  // inline double sumWKL(double s, double t) {
  //  int k,l;
  //  double sum_w_ab = 0.0; 
  //  for (k = 0; k <= 3; k++) {
  //    for (l = 0; l <=3; l++) {
  //      sum_w_ab += w(k, l, s, t);
  //    }
  //  }
  //  return sum_w_ab;
  //}
  
  // -------------------------------------------------------------------------
  // Refinement (The Oslo algorithm)
  // Generic implementation for functional and parametric case
  // There are 4 different configurations of vertices in the refined lattice:
  
  // For the uniform bicubic C2 case
  
  // 1) On an existing vertex
  template <class Type>
    inline Type phi_2i_2j(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/64.0*( mat(i-1,j-1) + mat(i-1,j+1) + mat(i+1,j-1) + mat(i+1,j+1)
      +6.0*(mat(i-1,j) + mat(i,j-1) + mat(i,j+1) + mat(i+1,j)) + 36.0*mat(i,j) );
  }
  
  // 2) On an existing vertical line
  template <class Type>
    inline Type phi_2i_2jPluss1(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/16.0*( mat(i-1,j) + mat(i-1,j+1) + mat(i+1,j) + mat(i+1,j+1)
      +6.0*(mat(i,j) + mat(i,j+1)) );
  }
  
  // 3) On an existing horizontal line
  template <class Type>
    inline Type phi_2iPlus1_2j(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/16.0*( mat(i,j-1) + mat(i,j+1) + mat(i+1,j-1) + mat(i+1,j+1)
      +6.0*(mat(i,j) + mat(i+1,j)) );
  }
  
  // 4) a new crossing in both directions
  template <class Type>
    inline Type phi_2iPlus1_2jPlus1(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/4.0*( mat(i,j) + mat(i,j+1) + mat(i+1,j) + mat(i+1,j+1) );
  }
  
  
  // The Oslo Algorithm For the uniform bicubic C1 case 
  // There are 16 configurations:
  // ----------------------------
  //
  //      |       :
  //      |       :
  //      |       :
  //.....9|13...15:4...........
  //     7|11    3:16
  //      |       :
  //      |       :
  //_____5|2____12:14__________
  //     1|6     8:10
  //      |       :
  //      |       :
  //      |       :
  
  // !!!! int i, int j denotes (2i) and (2j) in the coefficient matrix PHI !!!
  // 1) SW of an existing grid point
  template <class Type>
    inline Type phi_4iMin1_4jMin1(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/16.0*( 9.0*mat(i-1,j-1) + 3.0*(mat(i-1,j) + mat(i,j-1)) + mat(i,j) );
  }
  
  // 2) NE of an existing grid point
  template <class Type>
    inline Type phi_4i_4j(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/16.0*( mat(i-1,j-1) + 3.0*(mat(i-1,j) + mat(i,j-1)) + 9.0*mat(i,j) );
  }
  
  // 3) SW of a new grid point
  template <class Type>
    inline Type phi_4iPlus1_4jPlus1(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/64.0*( mat(i-1,j-1) + 5.0*(mat(i-1,j) + mat(i,j-1)) + 25.0*mat(i,j) + 2.0*(mat(i-1,j+1) + mat(i+1,j-1)) + 10.0*(mat(i+1,j) + mat(i,j+1)) + 4.0*mat(i+1,j+1) );
  }
  
  // 4) NE of a new grid point
  template <class Type>
    inline Type phi_4iPlus2_4jPlus2(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/64.0*( 4.0*mat(i,j) + 5.0*mat(i+2,j+1) + 2*(mat(i+2,j) + mat(i,j+2)) + mat(i+2,j+2) + 25*mat(i+1,j+1) + 10.0*(mat(i+1,j) + mat(i,j+1)) + 5.0*mat(i+1,j+2) );
  }
  
  // 5) NW of an existing grid point
  template <class Type>
    inline Type phi_4iMin1_4j(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/16.0*( 3.0*(mat(i-1,j-1) + mat(i,j)) + 9.0*mat(i-1,j) + mat(i,j-1) );
  }
  
  // 6) SE of an existing grid point
  template <class Type>
    inline Type phi_4i_4jMin1(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/16.0*( 3.0*(mat(i-1,j-1) + mat(i,j)) + mat(i-1,j) + 9.0*mat(i,j-1) );
  }
  
  // 7) SW of a new grid point on an existing vertical line
  template <class Type>
    inline Type phi_4iMin1_4jPlus1(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/32.0*( 3.0*mat(i-1,j-1) + 15.0*mat(i-1,j) + mat(i,j-1) + 5.0*mat(i,j) + 2.0*mat(i,j+1) + 6.0*mat(i-1,j+1) );
  }
  
  // 8) SW of a new grid point on an existing horizontal line
  template <class Type>
    inline Type phi_4iPlus1_4jMin1(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/32.0*( 3.0*mat(i-1,j-1) + mat(i-1,j) + 15*mat(i,j-1) + 5.0*mat(i,j) + 2.0*mat(i+1,j) + 6.0*mat(i+1,j-1) );
  }
  
  // 9) NW of a new grid point on an existing vertical line
  template <class Type>
    inline Type phi_4iMin1_4jPlus2(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/32.0*( 6.0*mat(i-1,j) + 2.0*mat(i,j) + 15*mat(i-1,j+1) + 5.0*mat(i,j+1) + mat(i,j+2) + 3.0*mat(i-1,j+2) );
  }
  
  // 10) SE of a new grid point on an existing horizontal line
  template <class Type>
    inline Type phi_4iPlus2_4jMin1(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/32.0*( 6.0*mat(i,j-1) + 2.0*mat(i,j) + mat(i+2,j) + 15.0*mat(i+1,j-1) + 5.0*mat(i+1,j) + 3.0*mat(i+2,j-1) );
  }
  
  // 11) SE of a new grid point on an existing vertical line
  template <class Type>
    inline Type phi_4i_4jPlus1(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/32.0*( mat(i-1,j-1) + 5.0*mat(i-1,j) + 3.0*mat(i,j-1) + 15.0*mat(i,j) + 2.0*mat(i-1,j+1) + 6.0*mat(i,j+1) );
  }
  
  // 12) NW of a new grid point on an existing horizontal line
  template <class Type>
    inline Type phi_4iPlus1_4j(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/32.0*( mat(i-1,j-1) + 3.0*mat(i-1,j) + 5.0*mat(i,j-1) + 15.0*mat(i,j) + 2.0*mat(i+1,j-1) + 6.0*mat(i+1,j) );
  }
  
  // 13) NE of a new grid point on an existing vertical line
  template <class Type>
    inline Type phi_4i_4jPlus2(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/32.0*( 2.0*mat(i-1,j) + 6.0*mat(i,j) + 5.0*mat(i-1,j+1) + 15.0*mat(i,j+1) + mat(i-1,j+2) + 3.0*mat(i,j+2) );
  }
  
  // 14) NE of a new grid point on an existing horizontal line
  template <class Type>
    inline Type phi_4iPlus2_4j(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/32.0*( 2.0*mat(i,j-1) + 6.0*mat(i,j) + 5.0*mat(i+1,j-1) + 15.0*mat(i+1,j) + mat(i+2,j-1) + 3.0*mat(i+2,j) );
  }
  
  // 15) NW of a new grid point
  template <class Type>
    inline Type phi_4iPlus1_4jPlus2(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/64.0*( 2.0*(mat(i-1,j) + mat(i+1,j+2)) + 10.0*(mat(i,j) + mat(i+1,j+1)) + 5.0*(mat(i-1,j+1) + mat(i,j+2)) + 25.0*mat(i,j+1) + 4.0*mat(i+1,j) + mat(i-1,j+2) );
  }
  
  // 16) SE of a new grid point
  template <class Type>
    inline Type phi_4iPlus2_4jPlus1(const GenMatrix<Type>& mat, int i, int j) {
    return 1.0/64.0*( 2.0*(mat(i,j-1) + mat(i+2,j+1)) + 10.0*(mat(i,j) + mat(i+1,j+1)) + 5.0*(mat(i+1,j-1) + mat(i+2,j)) + 4.0*mat(i,j+1) + 25.0*mat(i+1,j) + mat(i+2,j-1) );
  }
  
  // Refinement (Similar to the Oslo algorithm)
  void refineCoeffsC1(const GenMatrix<UCBspl_real>& PSI, GenMatrix<UCBspl_real>& PSIprime);
  void refineCoeffsC2(const GenMatrix<UCBspl_real>& PSI, GenMatrix<UCBspl_real>& PSIprime);

  // Restriction (This is not used by Multilvel B-splines) "Transposed" of the refinement operator
  bool restrictCoeffsC2(const GenMatrix<UCBspl_real>& rr, GenMatrix<UCBspl_real>& r);

}; // end namespace

#endif  

---