Merge cgogn:~vanhoey/CGoGN

Conflicts:
	CMakeLists.txt

README_ECLIPSE.TXT

-Pour utiliser Eclipse (testé avec la version Helios, Galileo & Indigo)
-English version bellow
+ *** English version below ***
 
+Pour utiliser Eclipse (testé avec la version Helios, Galileo, Indigo & Juno)
  * Configurer un projet dans Eclipse :
  - Faire un nouveau projet C++
  - Choisir Makefile project -> Empty Project
@@ -31,8 +31,8 @@ Projet X -> Properties -> Project references : cocher le projet dont il dépend.
 Afin d'éviter qu'Eclipse ne rejette les syntaxes Qt (les includes de Qt, les Q_OBJECT, les callbacks) :
  - Télécharger le plugin Qt "eclipse-integration" http://qt.nokia.com/developer/eclipse-integration/
  - Décompresser l'archive et copier son contenu dans le répertoire d'installation d'Eclipse (plugins et features)
- - Relancer Eclipse au moins une fois avec l'option --clean (relecture des répertoires locaux)
-Normalement, dans Window->Preferences de Eclipse, Qt apparait dans le menu.
+ - Relancer Eclipse au moins une fois avec l'option "-clean" (relecture des répertoires locaux)
+Normalement, dans Window --> Preferences de Eclipse, Qt apparait dans le menu.
  - Y ajouter le répertoire de Qt (/usr/bin sur Debian/Ubuntu) et le répertoire d'include (/usr/include/qt4).
 Ensuite, pour chaque projet, ajouter les includes suivants dans Properties -> C/C++ general -> Paths and Symbols
 parmi les includes de C++ :
@@ -63,7 +63,7 @@ Dans Preferences -> Shaders Preferences, on peut désormais configurer certaines
                                 ENGLISH VERSION
 =========================================================================================
 
-To use Eclipse (tested with version Helios, Galileo & Indigo)
+To use Eclipse (tested with version Helios, Galileoi, Indigo & Juno)
 
  * Setting up a project in Eclipse:
  - Make a new C + + project
@@ -84,18 +84,18 @@ In the project properties:
 	Tutorial: Build / Apps / Tutorial (compiles tutorials)
 	etc. ..
 
-This added to the directory of CGoGN a .project. and .cproject files.
+This adds a .project. and .cproject files to the CGoGN directory.
 
- * Inter-dependencies projects and autocompletion
-If the project X depends on another project, then:
+ * Project inter-dependencies and autocompletion
+If project X depends on another project, then:
 Project X -> Properties -> Project references: check the project which it depends.
 
  * Qt dependencies:
-To avoid Eclipse rejects the Qt syntax (the Qt includes the Q_OBJECT, callbacks):
+To avoid Eclipse rejecting the Qt syntax (includes, Q_OBJECT, callbacks):
  - Download the Qt plugin "eclipse-integration" http://qt.nokia.com/developer/eclipse-integration/
  - Unzip the archive and copy its contents into the installation directory of Eclipse (plugins and features)
- - Restart Eclipse at least once with the - clean (proofreading local directories)
-Normally in Window-> Preferences of Eclipse, Qt appears in the menu.
+ - Restart Eclipse at least once with the "-clean" (proofreading local directories)
+Normally in Window --> Preferences of Eclipse, Qt appears in the menu.
  - Add the Qt directory (/ usr / bin on Debian / Ubuntu) and the directory include (/ usr/include/qt4).
 Then, for each project, add the following includes in Properties -> C / C + + general -> Paths and Symbols
 among the C + + includes:

include/Algo/Decimation/approximator.h

@@ -39,14 +39,17 @@ namespace Decimation
 enum ApproximatorType
 {
 	A_QEM,
-	A_QEMhalfEdge,
 	A_MidEdge,
-	A_HalfCollapse,
 	A_CornerCutting,
 	A_TangentPredict1,
 	A_TangentPredict2,
-	A_LightfieldHalf,
-	A_LightfieldFull
+	A_ColorNaive,
+	A_ColorQEMext,
+	A_Lightfield,
+	// note: the following "h" prefix means that half-edges are prioritized instead of edges.
+	A_hHalfCollapse,
+	A_hQEM
+	// A_hLightfieldHalf,
 } ;
 
 template <typename PFP>
@@ -65,8 +68,10 @@ public:
 	{}
 	virtual ~ApproximatorGen()
 	{}
-	virtual const std::string& getApproximatedAttributeName() const = 0 ;
+	virtual const std::string& getApproximatedAttributeName(unsigned int index = 0) const = 0 ;
+//	virtual std::vector<std::string> getApproximatedAttributeNames() const = 0 ;
 	virtual ApproximatorType getType() const = 0 ;
+	virtual unsigned int getNbApproximated() const = 0 ;
 	virtual bool init() = 0 ;
 	virtual void approximate(Dart d) = 0 ;
 	virtual void saveApprox(Dart d) = 0 ;
@@ -87,52 +92,95 @@ public:
 protected:
 	Predictor<PFP, T>* m_predictor ;
 
-	VertexAttribute<T>& m_attrV ;	// vertex attribute to be approximated
-	EdgeAttribute<T> m_approx ;	// attribute to store approximation result
-	EdgeAttribute<T> m_detail ;	// attribute to store detail information for reconstruction
-	T m_app ;
+	std::vector<VertexAttribute<T>* > m_attrV ;	// vertex attributes to be approximated
+	std::vector<EdgeAttribute<T> > m_approx ;	// attributes to store approximation result
+	std::vector<EdgeAttribute<T> > m_detail ;	// attributes to store detail information for reconstruction
+	std::vector<T> m_app ;
 
 public:
-	Approximator(MAP& m, VertexAttribute<T>& a, Predictor<PFP, T>* predictor) :
-		ApproximatorGen<PFP>(m), m_predictor(predictor), m_attrV(a)
+	Approximator(MAP& m, std::vector<VertexAttribute<T>* > va, Predictor<PFP, T> * predictor) :
+		ApproximatorGen<PFP>(m), m_predictor(predictor), m_attrV(va)
 	{
-		std::stringstream aname ;
-		aname << "approx_" << m_attrV.name() ;
-		m_approx = this->m_map.template addAttribute<T, EDGE>(aname.str()) ;
-
-		if(m_predictor)	// if a predictor is associated to the approximator
-		{				// create an attribute to store the detail needed for reconstruction
-			std::stringstream dname ;
-			dname << "detail_" << m_attrV.name() ;
-			m_detail = this->m_map.template addAttribute<T, EDGE>(dname.str()) ;
+		const unsigned int& size = m_attrV.size() ;
+		assert(size > 0 || !"Approximator: no attributes provided") ;
+
+		m_approx.resize(size) ;
+		m_detail.resize(size) ;
+		m_app.resize(size) ;
+
+		for (unsigned int i = 0 ; i < size ; ++i)
+		{
+			if (!m_attrV[i]->isValid())
+				std::cerr << "Approximator Warning: attribute number " << i << " is not valid" << std::endl ;
+
+			std::stringstream aname ;
+			aname << "approx_" << m_attrV[i]->name() ;
+			m_approx[i] = this->m_map.template addAttribute<T, EDGE>(aname.str()) ;
+
+			if(m_predictor)	// if predictors are associated to the approximator
+			{				// create attributes to store the details needed for reconstruction
+				std::stringstream dname ;
+				dname << "detail_" << m_attrV[i]->name() ;
+				m_detail[i] = this->m_map.template addAttribute<T, EDGE>(dname.str()) ;
+			}
 		}
 	}
 
 	virtual ~Approximator()
 	{
-		this->m_map.template removeAttribute(m_approx) ;
-		if(m_predictor)
-			this->m_map.template removeAttribute(m_detail) ;
+		for (unsigned int i = 0 ; i < m_attrV.size() ; ++i)
+		{
+			this->m_map.template removeAttribute(m_approx[i]) ;
+			if(m_predictor)
+				this->m_map.template removeAttribute(m_detail[i]) ;
+		}
+	}
+
+	const std::string& getApproximatedAttributeName(unsigned int index = 0) const
+	{
+		return m_attrV[index]->name() ;
 	}
 
-	const std::string& getApproximatedAttributeName() const
+//	std::vector<std::string> getApproximatedAttributeNames() const
+//	{
+//		std::vector<std::string> names ;
+//		names.resize(m_attrV.size()) ;
+//		for (unsigned int i = 0 ; i < m_attrV.size() ; ++i)
+//			names[i] = m_attrV[i]->name() ;
+//
+//		return names ;
+//	}
+
+	unsigned int getNbApproximated() const
 	{
-		return m_attrV.name() ;
+		return m_attrV.size() ;
 	}
 
 	void saveApprox(Dart d)
 	{
-		m_app = m_approx[d] ;
+		for (unsigned int i = 0 ; i < m_attrV.size() ; ++i)
+			m_app[i] = m_approx[i][d] ;
 	}
 
 	void affectApprox(Dart d)
 	{
-		m_attrV[d] = m_app ;
+		for (unsigned int i = 0 ; i < m_attrV.size() ; ++i)
+			m_attrV[i]->operator[](d) = m_app[i] ;
+	}
+
+	const T& getApprox(Dart d, unsigned int index = 0) const
+	{
+		return m_approx[index][d] ;
 	}
 
-	const T& getApprox(Dart d) const
+	std::vector<T> getAllApprox(Dart d) const
 	{
-		return m_approx[d] ;
+		std::vector<T> res ;
+		res.resize(m_attrV.size()) ;
+		for (unsigned int i = 0 ; i < m_attrV.size() ; ++i)
+			res[i] = m_approx[i][d] ;
+
+		return res ;
 	}
 
 	const Predictor<PFP, T>* getPredictor() const
@@ -140,18 +188,38 @@ public:
 		return m_predictor ;
 	}
 
-	const T& getDetail(Dart d) const
+	const T& getDetail(Dart d, unsigned int index = 0) const
+	{
+		assert(m_predictor || !"Trying to get detail on a non-predictive scheme") ;
+		return m_detail[index][d] ;
+	}
+
+	std::vector<T> getAllDetail(Dart d) const
 	{
 		assert(m_predictor || !"Trying to get detail on a non-predictive scheme") ;
-		return m_detail[d] ;
+
+		std::vector<T> res ;
+		res.resize(m_attrV.size()) ;
+		for (unsigned int i = 0 ; i < m_attrV.size() ; ++i)
+			res[i] = m_detail[i][d] ;
+		return res ;
+	}
+
+	void setDetail(Dart d, unsigned int index, T& val)
+	{
+		assert(m_predictor || !"Trying to set detail on a non-predictive scheme") ;
+		m_detail[index][d] = val ;
 	}
 
-	void setDetail(Dart d, T& val)
+	void setDetail(Dart d, std::vector<T>& val)
 	{
 		assert(m_predictor || !"Trying to set detail on a non-predictive scheme") ;
-		m_detail[d] = val ;
+
+		for (unsigned int i = 0 ; i < m_attrV.size() ; ++i)
+			m_detail[index][d] = val[i] ;
 	}
 
+
 //	// TODO works only for vector types !!
 //	REAL detailMagnitude(Dart d)
 //	{
@@ -162,13 +230,17 @@ public:
 	void addDetail(Dart d, double amount, bool sign, typename PFP::MATRIX33* detailTransform)
 	{
 		assert(m_predictor || !"Trying to add detail on a non-predictive scheme") ;
-		T det = m_detail[d] ;
-		if(detailTransform)
-			det = (*detailTransform) * det ;
-		det *= amount ;
-		if(!sign)
-			det *= REAL(-1) ;
-		m_attrV[d] += det ;
+
+		for (unsigned int i = 0 ; i < m_attrV.size() ; ++i)
+		{
+			T det = m_detail[i][d] ;
+			if(detailTransform)
+				det = (*detailTransform) * det ;
+			det *= amount ;
+			if(!sign)
+				det *= REAL(-1) ;
+			m_attrV[i]->operator[](d) += det ;
+		}
 	}
 } ;
 

include/Algo/Decimation/colorPerVertexApproximator.h

+/*******************************************************************************
+* CGoGN: Combinatorial and Geometric modeling with Generic N-dimensional Maps  *
+* version 0.1                                                                  *
+* Copyright (C) 2009-2012, IGG Team, LSIIT, University of Strasbourg           *
+*                                                                              *
+* This library is free software; you can redistribute it and/or modify it      *
+* under the terms of the GNU Lesser General Public License as published by the *
+* Free Software Foundation; either version 2.1 of the License, or (at your     *
+* option) any later version.                                                   *
+*                                                                              *
+* This library 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 Lesser General Public License  *
+* for more details.                                                            *
+*                                                                              *
+* You should have received a copy of the GNU Lesser General Public License     *
+* along with this library; if not, write to the Free Software Foundation,      *
+* Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301 USA.           *
+*                                                                              *
+* Web site: http://cgogn.unistra.fr/                                           *
+* Contact information: cgogn@unistra.fr                                        *
+*                                                                              *
+*******************************************************************************/
+
+#ifndef __COLOR_APPROXIMATOR_H__
+#define __COLOR_APPROXIMATOR_H__
+
+#include "Algo/Decimation/approximator.h"
+#include "Topology/generic/mapBrowser.h"
+
+namespace CGoGN
+{
+
+namespace Algo
+{
+
+namespace Decimation
+{
+
+template <typename PFP>
+class Approximator_ColorNaive : public Approximator<PFP, typename PFP::VEC3>
+{
+public:
+	typedef typename PFP::MAP MAP ;
+	typedef typename PFP::VEC3 VEC3 ;
+	typedef typename PFP::REAL REAL ;
+
+protected:
+	VertexAttribute<VEC3> m_position ;
+	EdgeAttribute<VEC3> m_approxposition ;
+
+	VertexAttribute<VEC3> *m_color ;
+
+public:
+	Approximator_ColorNaive(MAP& m, std::vector<VertexAttribute<VEC3>* >& attr, Predictor<PFP, VEC3>* pred = NULL) :
+		Approximator<PFP, VEC3>(m, attr, pred)
+	{
+		m_color = this->m_attrV[0] ;
+	}
+	~Approximator_ColorNaive()
+	{}
+
+	ApproximatorType getType() const
+	{
+		return A_ColorNaive ;
+	}
+
+	bool init()
+	{
+		assert(m_color->isValid() || !"Approximator_ColorNaive: the approximated attribute is not valid") ;
+
+		m_position = this->m_map.template getAttribute<VEC3, VERTEX>("position") ;
+		assert(m_position.isValid() || !"Approximator_ColorNaive::init: the position attribute is not valid") ;
+
+		m_approxposition = this->m_map.template getAttribute<VEC3, EDGE>("approx_position") ;
+		assert(m_approxposition.isValid() || !"Approximator_ColorNaive::init: the approx_position attribute is not valid") ;
+
+		return m_color->isValid() && m_position.isValid() && m_approxposition.isValid() ;
+	}
+
+	void approximate(Dart d) ;
+} ;
+
+template <typename PFP>
+class Approximator_ColorQEMext : public Approximator<PFP, typename PFP::VEC3>
+{
+public:
+	typedef typename PFP::MAP MAP ;
+	typedef typename PFP::REAL REAL ;
+	typedef typename PFP::VEC3 VEC3 ;
+	typedef Geom::Vector<6,REAL> VEC6 ;
+
+protected:
+	VertexAttribute<QuadricNd<REAL,6> > m_quadric ;
+	VertexAttribute<VEC3> *m_position ;
+	VertexAttribute<VEC3> *m_color ;
+
+public:
+	Approximator_ColorQEMext(MAP& m, std::vector<VertexAttribute<VEC3>* >& attr, Predictor<PFP, VEC3>* pred = NULL) :
+		Approximator<PFP, VEC3>(m, attr, pred)
+	{
+		assert(attr.size() > 1 || !"Approximator_ColorQEMext: there are not sufficient attributes provided") ;
+
+		m_position = this->m_attrV[0] ;
+		m_color = this->m_attrV[1] ;
+	}
+
+	~Approximator_ColorQEMext()
+	{}
+
+	ApproximatorType getType() const
+	{
+		return A_ColorQEMext ;
+	}
+
+	bool init() ;
+
+	void approximate(Dart d) ;
+} ;
+
+} //namespace Decimation
+
+} //namespace Algo
+
+} //namespace CGoGN
+
+#include "Algo/Decimation/colorPerVertexApproximator.hpp"
+
+#endif

include/Algo/Decimation/colorPerVertexApproximator.hpp

+/*******************************************************************************
+* CGoGN: Combinatorial and Geometric modeling with Generic N-dimensional Maps  *
+* version 0.1                                                                  *
+* Copyright (C) 2009-2012, IGG Team, LSIIT, University of Strasbourg           *
+*                                                                              *
+* This library is free software; you can redistribute it and/or modify it      *
+* under the terms of the GNU Lesser General Public License as published by the *
+* Free Software Foundation; either version 2.1 of the License, or (at your     *
+* option) any later version.                                                   *
+*                                                                              *
+* This library 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 Lesser General Public License  *
+* for more details.                                                            *
+*                                                                              *
+* You should have received a copy of the GNU Lesser General Public License     *
+* along with this library; if not, write to the Free Software Foundation,      *
+* Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301 USA.           *
+*                                                                              *
+* Web site: http://cgogn.unistra.fr/                                           *
+* Contact information: cgogn@unistra.fr                                        *
+*                                                                              *
+*******************************************************************************/
+
+namespace CGoGN
+{
+
+namespace Algo
+{
+
+namespace Decimation
+{
+
+/************************************************************************************
+ *                      NAIVE COLOR METRIC                                          *
+ ************************************************************************************/
+
+template <typename PFP>
+void Approximator_ColorNaive<PFP>::approximate(Dart d)
+{
+	Dart dd = this->m_map.phi1(d) ;
+
+	const VEC3& p1 = m_position.operator[](d) ;
+	const VEC3& p2 = m_position.operator[](dd) ;
+	const VEC3& p = m_approxposition.operator[](d) ;
+
+	const VEC3& edge = p2 - p1 ;
+	const REAL& ratio = std::max(std::min(((p - p1) * edge) / edge.norm2(),REAL(1)),REAL(0)) ;
+
+	this->m_approx[0][d] = m_color->operator[](d)*ratio + m_color->operator[](dd)*(1-ratio) ;
+}
+
+/************************************************************************************
+ *                            EXTENDED QUADRIC ERROR METRIC                         *
+ ************************************************************************************/
+template <typename PFP>
+bool Approximator_ColorQEMext<PFP>::init()
+{
+	m_quadric = this->m_map.template getAttribute<QuadricNd<REAL,6>, VERTEX>("QEMext-quadric") ;
+	// Does not require to be valid (if it is not, altenatives will be used).
+
+	if(this->m_predictor)
+	{
+		return false ;
+	}
+
+	return m_position->isValid() && m_color->isValid() ;
+}
+
+template <typename PFP>
+void Approximator_ColorQEMext<PFP>::approximate(Dart d)
+{
+	MAP& m = this->m_map ;
+
+	// get some darts
+	Dart dd = m.phi2(d) ;
+
+	QuadricNd<REAL,6> q1, q2 ;
+	if(!m_quadric.isValid()) // if the selector is not QEM, compute local error quadrics
+	{
+		// compute the error quadric associated to v1
+		Dart it = d ;
+		do
+		{
+			VEC6 p0,p1,p2 ;
+			for (unsigned int i = 0 ; i < 3 ; ++i)
+			{
+				p0[i] = this->m_attrV[0]->operator[](it)[i] ;
+				p0[i+3] = this->m_attrV[1]->operator[](it)[i] ;
+				p1[i] = this->m_attrV[0]->operator[](m.phi1(it))[i] ;
+				p1[i+3] = this->m_attrV[1]->operator[](m.phi1(it))[i] ;
+				p2[i] = this->m_attrV[0]->operator[](m.phi_1(it))[i] ;
+				p2[i+3] = this->m_attrV[1]->operator[](m.phi_1(it))[i] ;
+			}
+
+			QuadricNd<REAL,6> q(p0,p1,p2) ;
+			q1 += q ;
+			it = m.phi2_1(it) ;
+		} while(it != d) ;
+
+		// compute the error quadric associated to v2
+		it = dd ;
+		do
+		{
+			VEC6 p0,p1,p2 ;
+			for (unsigned int i = 0 ; i < 3 ; ++i)
+			{
+				p0[i] = this->m_attrV[0]->operator[](it)[i] ;
+				p0[i+3] = this->m_attrV[1]->operator[](it)[i] ;
+				p1[i] = this->m_attrV[0]->operator[](m.phi1(it))[i] ;
+				p1[i+3] = this->m_attrV[1]->operator[](m.phi1(it))[i] ;
+				p2[i] = this->m_attrV[0]->operator[](m.phi_1(it))[i] ;
+				p2[i+3] = this->m_attrV[1]->operator[](m.phi_1(it))[i] ;
+			}
+
+			QuadricNd<REAL,6> q(p0,p1,p2) ;
+			q2 += q ;
+			it = m.phi2_1(it) ;
+		} while(it != dd) ;
+	}
+	else // if the selector is QEM, use the error quadrics computed by the selector
+	{
+		q1 = m_quadric[d] ;
+		q2 = m_quadric[dd] ;
+	}
+
+	QuadricNd<REAL,6> quad ;
+	quad += q1 ;	// compute the sum of the
+	quad += q2 ;	// two vertices quadrics
+
+	VEC6 res ;
+	bool opt = quad.findOptimizedVec(res) ;	// try to compute an optimized position for the contraction of this edge
+	if(!opt)
+	{
+		VEC6 p1, p2 ;
+		for (unsigned int i = 0 ; i < 3; ++i)
+		{
+			p1[i] = this->m_attrV[0]->operator[](d)[i] ;	// let the new vertex lie
+			p1[i+3] = this->m_attrV[1]->operator[](d)[i] ; // on either one
+			p2[i] = this->m_attrV[0]->operator[](dd)[i] ;	// of the two
+			p2[i+3] = this->m_attrV[1]->operator[](dd)[i] ; // endpoints
+		}
+		VEC6 p12 = (p1 + p2) / 2.0f ;	// or the middle of the edge
+
+		REAL e1 = quad(p1) ;
+		REAL e2 = quad(p2) ;
+		REAL e12 = quad(p12) ;
+		REAL minerr = std::min(std::min(e1, e2), e12) ;	// consider only the one for
+		if(minerr == e12)
+			res = p12 ;		// which the error is minimal
+		else if(minerr == e1)
+			res = p1 ;
+		else
+			res = p2 ;
+	}
+
+	// copy res into m_approx
+	for (unsigned int i = 0 ; i < 3 ; ++i)
+	{
+		this->m_approx[0][d][i] = res[i] ;
+		this->m_approx[1][d][i] = res[i+3] ;
+	}
+}
+
+} //namespace Decimation
+
+} //namespace Algo
+
+} //namespace CGoGN

include/Algo/Decimation/decimation.h

@@ -28,6 +28,7 @@
 #include "Algo/Decimation/edgeSelector.h"
 #include "Algo/Decimation/halfEdgeSelector.h"
 #include "Algo/Decimation/geometryApproximator.h"
+#include "Algo/Decimation/colorPerVertexApproximator.h"
 #include "Algo/Decimation/lightfieldApproximator.h"
 
 namespace CGoGN
@@ -44,9 +45,10 @@ void decimate(
 	typename PFP::MAP& map,
 	SelectorType s,
 	ApproximatorType a,
-	VertexAttribute<typename PFP::VEC3>& position,
+	std::vector<VertexAttribute<typename PFP::VEC3> *>& position,
 	unsigned int nbWantedVertices,
-	const FunctorSelect& selected = allDarts
+	const FunctorSelect& selected = allDarts,
+	void (*callback_wrapper)(void*, const void*) = NULL, void *callback_object = NULL
 ) ;
 
 } //namespace Decimation

include/Algo/Decimation/decimation.hpp

@@ -34,75 +34,106 @@ namespace Decimation
 template <typename PFP>
 void decimate(
 	typename PFP::MAP& map, SelectorType s, ApproximatorType a,
-	VertexAttribute<typename PFP::VEC3>& position, unsigned int nbWantedVertices, const FunctorSelect& selected
+	std::vector<VertexAttribute<typename PFP::VEC3>* >& attribs, unsigned int nbWantedVertices, const FunctorSelect& selected,
+	void (*callback_wrapper)(void*, const void*), void* callback_object
 )
 {
+	assert(attribs.size() >= 1 || !"Decimate: not enough attribs provided") ;
+	assert(attribs[0]->name() == "position" || !"Decimate: first attribute is not position") ;
+	VertexAttribute<typename PFP::VEC3> position = *(attribs[0]) ;
+
 	std::vector<ApproximatorGen<PFP>*> approximators ;
 	EdgeSelector<PFP>* selector = NULL ;