Skip to content
GitLab
Commit 0f99fbb3 authored by Laurent Belcour's avatar Laurent Belcour
Browse files

Cleaning quadprog and matlab code to obtain the same thing

parent 3801ebe2
Hide whitespace changes
Inline Side-by-side
sources/core/function.cpp
View file @ 0f99fbb3
......@@ -48,10 +48,11 @@ vec function::getMax() const
void function::save(const std::string& filename, const arguments& args) const
{
bool is_cpp = args["export"] == "C++";
bool is_explorer = args["export"] == "explorer";
bool is_shader = args["export"] == "shader" || is_explorer;
bool is_matlab = args["export"] == "matlab";
bool is_alta = !args.is_defined("export") || args["export"] == "alta";
bool is_cpp = args["export"] == "C++";
bool is_explorer = args["export"] == "explorer";
bool is_shader = args["export"] == "shader" || is_explorer;
bool is_matlab = args["export"] == "matlab";
// Open the file
std::ofstream file(filename.c_str());
......@@ -60,60 +61,66 @@ void function::save(const std::string& filename, const arguments& args) const
std::cerr << "<<ERROR>> unable to open output file for writing" << std::endl;
}
if(is_explorer)
{
file << "analytic" << std::endl;
file << std::endl;
file << "::begin shader" << std::endl;
}
// If the export is the alta format, use the maximum precision formatting
if(is_alta)
{
file.precision(10);
}
if(is_explorer)
{
file << "analytic" << std::endl;
file << std::endl;
file << "::begin shader" << std::endl;
}
// Save common header
save_header(file, args);
save_header(file, args);
// Save function definition
save_body(file, args);
save_body(file, args);
if(is_cpp)
{
file << "vec brdf(const vec& in, const vec& file)" << std::endl;
file << "{" << std::endl;
file << "\tvec res(" << dimY() << ");" << std::endl;
file << "\t";
}
else if(is_matlab)
{
file << "function res = brdf(in, file)" << std::endl;
file << "{" << std::endl;
file << "\tres = zeros(" << dimY() << ");" << std::endl;
file << "\t";
}
else if(is_shader)
{
file << "vec3 BRDF(vec3 L, vec3 V, vec3 N, vec3 X, vec3 Y)" << std::endl;
file << "{" << std::endl;
file << "\tvec3 res = ";
}
if(is_cpp)
{
file << "vec brdf(const vec& in, const vec& file)" << std::endl;
file << "{" << std::endl;
file << "\tvec res(" << dimY() << ");" << std::endl;
file << "\t";
}
else if(is_matlab)
{
file << "function res = brdf(in, file)" << std::endl;
file << "{" << std::endl;
file << "\tres = zeros(" << dimY() << ");" << std::endl;
file << "\t";
}
else if(is_shader)
{
file << "vec3 BRDF(vec3 L, vec3 V, vec3 N, vec3 X, vec3 Y)" << std::endl;
file << "{" << std::endl;
file << "\tvec3 res = ";
}
// Save fit data
save_call(file, args);
if(is_cpp || is_shader)
{
file << ";" << std::endl;
file << "\treturn res;" << std::endl;
file << "}" << std::endl;
}
else if(is_matlab)
{
file << ";" << std::endl;
file << "\treturn res;" << std::endl;
file << "endfunction" << std::endl;
}
file << std::endl;
save_call(file, args);
if(is_cpp || is_shader)
{
file << ";" << std::endl;
file << "\treturn res;" << std::endl;
file << "}" << std::endl;
}
else if(is_matlab)
{
file << ";" << std::endl;
file << "\treturn res;" << std::endl;
file << "endfunction" << std::endl;
}
file << std::endl;
if(is_explorer)
{
file << "::end shader" << std::endl;
}
if(is_explorer)
{
file << "::end shader" << std::endl;
}
}
//! \brief save the header of the output function file. The header should
......@@ -128,9 +135,10 @@ void function::save_header(std::ostream& out, const arguments& args) const
out << "#DIM " << _nX << " " << _nY << std::endl;
out << "#PARAM_IN " << params::get_name(input_parametrization()) << std::endl;
//out << "#PARAM_OUT " << params::get_name(output_parametrization()) << std::endl;*
if(args.is_defined("export-append")) {
out << args["export-append"] << std::endl;
}
if(args.is_defined("export-append"))
{
out << args["export-append"] << std::endl;
}
out << "#ALTA HEADER END" << std::endl;
out << std::endl;
}
......
sources/core/plugins_manager.cpp
View file @ 0f99fbb3
......@@ -208,6 +208,9 @@ function* plugins_manager::get_function(const std::string& filename)
throw ;
}
// Set the precision of the input
file.precision(10);
// Parameters of the function object
int nX, nY;
arguments args;
......
sources/core/rational_function.cpp
View file @ 0f99fbb3
......@@ -35,7 +35,13 @@ void rational_function_1d::update(const vec& in_a,
a.resize(in_a.size()) ;
b.resize(in_b.size()) ;
const double b0 = 1.0;//in_b[0];
#define NORMALIZE
#ifdef NORMALIZE
const double b0 = (std::abs(in_b[0]) > 1.0E-10) ? in_b[0] : 1.0;
#else
const double b0 = 1.0;
#endif
for(int i=0; i<a.size(); ++i) { a[i] = in_a[i] / b0; }
for(int i=0; i<b.size(); ++i) { b[i] = in_b[i] / b0; }
......
sources/plugins/rational_fitter_matlab/rational_fitter.cpp
View file @ 0f99fbb3
......@@ -124,11 +124,7 @@ bool rational_fitter_matlab::fit_data(const vertical_segment* d, int np, int nq,
// Size of the problem
int N = np+nq ;
int M = d->size() ;
/*
// Get the maximum value in data to scale the input parameter space
// so that it reduces the values of the polynomial
vec dmax = d->max() ;
*/
// Matrices of the problem
Eigen::MatrixXd G (N, N) ;
Eigen::VectorXd g (N) ;
......@@ -171,26 +167,27 @@ bool rational_fitter_matlab::fit_data(const vertical_segment* d, int np, int nq,
if(j<np)
{
const double pi = r->p(xi, j) ;
a0_norm += pi*pi ;
a1_norm += pi*pi ;
// Updating Eigen matrix
CI(2*i+0, j) = pi ;
CI(2*i+1, j) = -pi ;
CI(2*i+0, j) = pi;
CI(2*i+1, j) = -pi;
}
// Filling the q part
else
{
vec yl, yu ;
d->get(i, yl, yu) ;
vec yl, yu;
d->get(i, yl, yu);
const double qi = r->q(xi, j-np) ;
a0_norm += qi*qi * (yu[ny]*yu[ny]) ;
a1_norm += qi*qi * (yl[ny]*yl[ny]) ;
const double qi = r->q(xi, j-np);
// Updating Eigen matrix
CI(2*i+0, j) = -yu[ny] * qi ;
CI(2*i+1, j) = yl[ny] * qi ;
CI(2*i+0, j) = -yu[ny] * qi;
CI(2*i+1, j) = yl[ny] * qi;
}
// Update the norm of the row
a0_norm += CI(2*i+0, j)*CI(2*i+0, j);
a1_norm += CI(2*i+1, j)*CI(2*i+1, j);
}
// Set the c vector, will later be updated using the
......
sources/plugins/rational_fitter_quadprog/rational_fitter.cpp
View file @ 0f99fbb3
......@@ -122,32 +122,25 @@ bool rational_fitter_quadprog::fit_data(const vertical_segment* d, int np, int n
// np and nq are the degree of the RP to fit to the data
// y is the dimension to fit on the y-data (e.g. R, G or B for RGB signals)
// the function return a ration BRDF function and a boolean
bool rational_fitter_quadprog::fit_data(const vertical_segment* dat, int np, int nq, int ny, rational_function_1d* r)
bool rational_fitter_quadprog::fit_data(const vertical_segment* d, int np, int nq, int ny, rational_function_1d* r)
{
const vertical_segment* d = dynamic_cast<const vertical_segment*>(dat) ;
if(r == NULL || d == NULL)
{
std::cerr << "<<ERROR>> not passing the correct class to the fitter" << std::endl ;
return false ;
}
// Get the maximum value in data to scale the input parameter space
// so that it reduces the values of the polynomial
vec dmax = d->max() ;
// Size of the problem
const int N = np+nq ;
const int M = d->size() ;
// Matrices of the problem
QuadProgPP::Matrix<double> G (0.0, np+nq, np+nq) ;
QuadProgPP::Vector<double> g (0.0, np+nq) ;
QuadProgPP::Matrix<double> CI(0.0, np+nq, 2*d->size()) ;
QuadProgPP::Vector<double> ci(0.0, 2*d->size()) ;
QuadProgPP::Matrix<double> CE(0.0, np+nq, 0) ;
QuadProgPP::Matrix<double> G (0.0, N, N) ;
QuadProgPP::Vector<double> g (0.0, N) ;
QuadProgPP::Matrix<double> CI(0.0, N, 2*M) ;
QuadProgPP::Vector<double> ci(0.0, 2*M) ;
QuadProgPP::Matrix<double> CE(0.0, N, 0) ;
QuadProgPP::Vector<double> ce(0.0, 0) ;
Eigen::MatrixXd eCI(np+nq, 2*d->size()) ;
Eigen::MatrixXd eCI(2*M, N) ;
// Select the size of the result vector to
// be equal to the dimension of p + q
for(int i=0; i<np+nq; ++i)
for(int i=0; i<N; ++i)
{
G[i][i] = 1.0 ;
}
......@@ -155,7 +148,7 @@ bool rational_fitter_quadprog::fit_data(const vertical_segment* dat, int np, int
// Each constraint (fitting interval or point
// add another dimension to the constraint
// matrix
for(int i=0; i<d->size(); ++i)
for(int i=0; i<M; ++i)
{
// Norm of the row vector
double a0_norm = 0.0 ;
......@@ -163,30 +156,23 @@ bool rational_fitter_quadprog::fit_data(const vertical_segment* dat, int np, int
vec xi = d->get(i) ;
bool is_boundary = false;
for(int k=0; k<d->dimX(); ++k)
{
is_boundary = is_boundary || (xi[k] <= d->min()[k]) || (xi[k] >= d->max()[k]);
}
// A row of the constraint matrix has this
// form: [p_{0}(x_i), .., p_{np}(x_i), -f(x_i) q_{0}(x_i), .., -f(x_i) q_{nq}(x_i)]
// For the lower constraint and negated for
// the upper constraint
for(int j=0; j<np+nq; ++j)
for(int j=0; j<N; ++j)
{
// Filling the p part
if(j<np)
{
const double pi = r->p(xi, j) ;
a0_norm += pi*pi ;
a1_norm += pi*pi ;
CI[j][i] = pi ;
CI[j][i+d->size()] = -pi ;
const double pi = r->p(xi, j);
CI[j][2*i+0] = pi;
CI[j][2*i+1] = -pi;
// Updating Eigen matrix
eCI(j,i) = pi ;
eCI(j,i+d->size()) = -pi ;
eCI(2*i+0, j) = CI[j][2*i+0];
eCI(2*i+1, j) = CI[j][2*i+1];
}
// Filling the q part
else
......@@ -194,67 +180,51 @@ bool rational_fitter_quadprog::fit_data(const vertical_segment* dat, int np, int
vec yl, yu ;
d->get(i, yl, yu) ;
// Add a constraints for boundary conditions
if(is_boundary)
{
vec mean = 0.5*(yl+yu);
yl = mean + _boundary * (yl - mean);
yu = mean + _boundary * (yu - mean);
}
const double qi = r->q(xi, j-np) ;
a0_norm += qi*qi * (yl[ny]*yl[ny]) ;
CI[j][i] = -yl[ny] * qi ;
a1_norm += qi*qi * (yu[ny]*yu[ny]) ;
CI[j][i+d->size()] = yu[ny] * qi ;
const double qi = r->q(xi, j-np);
CI[j][2*i+0] = -yu[ny] * qi;
CI[j][2*i+1] = yl[ny] * qi;
// Updating Eigen matrix
eCI(j,i) = -yl[ny] * qi ;
eCI(j,i+d->size()) = yu[ny] * qi ;
eCI(2*i+0, j) = CI[j][2*i+0];
eCI(2*i+1, j) = CI[j][2*i+1];
}
// Update the norm of the row
a0_norm += CI[j][2*i+0]*CI[j][2*i+0];
a1_norm += CI[j][2*i+1]*CI[j][2*i+1];
}
// Set the c vector, will later be updated using the
// delta parameter.
ci[i] = -sqrt(a0_norm) ;
ci[i+d->size()] = -sqrt(a1_norm) ;
ci[2*i+0] = -sqrt(a0_norm) ;
ci[2*i+1] = -sqrt(a1_norm) ;
}
#ifdef DEBUG
std::cout << "CI = [" ;
for(int j=0; j<d->size()*2; ++j)
for(int j=0; j<2*M; ++j)
{
for(int i=0; i<np+nq; ++i)
for(int i=0; i<N; ++i)
{
std::cout << CI[i][j] ;
if(i != np+nq-1) std::cout << ", ";
if(i != N-1) std::cout << ", ";
}
if(j != d->size()*2-1)
if(j != 2*M-1)
std::cout << ";" << std::endl;
else
std::cout << "]" << std::endl ;
}
#endif
/*
QTime time ;
time.start() ;
*/
// Update the ci column with the delta parameter
// (See Celis et al. 2007 p.12)
Eigen::JacobiSVD<Eigen::MatrixXd, Eigen::HouseholderQRPreconditioner> svd(eCI);
const double sigma_m = svd.singularValues()(std::min(2*d->size(), np+nq)-1) ;
const double sigma_m = svd.singularValues()(std::min(2*M, N)-1) ;
const double sigma_M = svd.singularValues()(0) ;
/*
int msec = time.elapsed() ;
int sec = (msec / 1000) % 60 ;
int min = (msec / 60000) % 60 ;
int hour = (msec / 3600000) ;
std::cout << "<<INFO>> delta took " << hour << "h " << min << "m " << sec << "s" << std::endl ;
*/
#ifdef DEBUG
std::cout << "<<DEBUG>> SVD = [ " ;
for(int i=0; i<std::min(2*d->size(), np+nq); ++i)
for(int i=0; i<std::min(2*M, N); ++i)
{
std::cout << svd.singularValues()(i) << ", " ;
}
......@@ -270,16 +240,16 @@ bool rational_fitter_quadprog::fit_data(const vertical_segment* dat, int np, int
#endif
return false ;
}
else if(delta <= 0.0)
else if(delta <= 0.0)
{
delta = 1.0 ;
}
#ifdef DEBUG
#ifndef DEBUG
std::cout << "<<DEBUG>> delta factor: " << sigma_m << " / " << sigma_M << " = " << delta << std::endl ;
#endif
for(int i=0; i<2*d->size(); ++i)
for(int i=0; i<2*M; ++i)
{
ci[i] = ci[i] * delta ;
#ifdef DEBUG
......@@ -308,7 +278,7 @@ bool rational_fitter_quadprog::fit_data(const vertical_segment* dat, int np, int
// Recopy the vector d
vec p(np), q(nq);
double norm = 0.0 ;
for(int i=0; i<np+nq; ++i)
for(int i=0; i<N; ++i)
{
const double v = x[i];
norm += v*v ;
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment