ErrorMinimizer.cpp

Go to the documentation of this file.

// kate: replace-tabs off; indent-width 4; indent-mode normal
// vim: ts=4:sw=4:noexpandtab
/*

Copyright (c) 2010--2012,
François Pomerleau and Stephane Magnenat, ASL, ETHZ, Switzerland
You can contact the authors at <f dot pomerleau at gmail dot com> and
<stephane at magnenat dot net>

All rights reserved.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
    * Redistributions of source code must retain the above copyright
      notice, this list of conditions and the following disclaimer.
    * Redistributions in binary form must reproduce the above copyright
      notice, this list of conditions and the following disclaimer in the
      documentation and/or other materials provided with the distribution.
    * Neither the name of the <organization> nor the
      names of its contributors may be used to endorse or promote products
      derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
DISCLAIMED. IN NO EVENT SHALL ETH-ASL BE LIABLE FOR ANY
DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
(INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

*/

#include "PointMatcher.h"
#include "PointMatcherPrivate.h"

template<typename T>
PointMatcher<T>::ErrorMinimizer::ErrorMinimizer():
    pointUsedRatio(-1.),
    weightedPointUsedRatio(-1.)
{}

template<typename T>
PointMatcher<T>::ErrorMinimizer::ErrorMinimizer(const std::string& className, const ParametersDoc paramsDoc, const Parameters& params):
    Parametrizable(className,paramsDoc,params)
{}

template<typename T>
PointMatcher<T>::ErrorMinimizer::~ErrorMinimizer()
{}

template<typename T>
PointMatcher<T>::ErrorMinimizer::ErrorElements::ErrorElements(const DataPoints& reading, const DataPoints reference, const OutlierWeights weights, const Matches matches):
    reading(reading),
    reference(reference),
    weights(weights),
    matches(matches)
{
    assert(reading.features.cols() == reference.features.cols());
    assert(reading.features.cols() == weights.cols());
    assert(reading.features.cols() == matches.dists.cols());
    // May have no descriptors... size 0
}

template<typename T>
T PointMatcher<T>::ErrorMinimizer::getPointUsedRatio() const
{
    return pointUsedRatio;
}

template<typename T>
T PointMatcher<T>::ErrorMinimizer::getWeightedPointUsedRatio() const
{
    return weightedPointUsedRatio;
}

template<typename T>
T PointMatcher<T>::ErrorMinimizer::getOverlap() const
{
    LOG_INFO_STREAM("ErrorMinimizer - warning, no specific method to compute overlap was provided for the ErrorMinimizer used.");
    return weightedPointUsedRatio;
}

template<typename T>
typename PointMatcher<T>::Matrix PointMatcher<T>::ErrorMinimizer::crossProduct(const Matrix& A, const Matrix& B)
{
    //Note: A = [x, y, z, 1] and B = [x, y, z] for convenience

    // Expecting matched points
    assert(A.cols() == B.cols());

    // Expecting homogenous coord X eucl. coord
    assert(A.rows() -1 == B.rows());

    // Expecting homogenous coordinates
    assert(A.rows() == 4 || A.rows() == 3);
    
    const unsigned int x = 0;
    const unsigned int y = 1;
    const unsigned int z = 2;

    Matrix cross;
    if(A.rows() == 4)
    {
        cross = Matrix(B.rows(), B.cols());
                
        cross.row(x) = A.row(y).cwise() * B.row(z) - A.row(z).cwise() * B.row(y);
        cross.row(y) = A.row(z).cwise() * B.row(x) - A.row(x).cwise() * B.row(z);
        cross.row(z) = A.row(x).cwise() * B.row(y) - A.row(y).cwise() * B.row(x);
    }
    else
    {
        //pseudo-cross product for 2D vectors
        cross = Vector(B.cols());
        cross = A.row(x).cwise() * B.row(y) - A.row(y).cwise() * B.row(x);
    }

    return cross;
}


template<typename T>
typename PointMatcher<T>::ErrorMinimizer::ErrorElements& PointMatcher<T>::ErrorMinimizer::getMatchedPoints(
        const DataPoints& requestedPts,
        const DataPoints& sourcePts,
        const Matches& matches, 
        const OutlierWeights& outlierWeights)
{
    typedef typename Matches::Ids Ids;
    typedef typename Matches::Dists Dists;
    
    assert(matches.ids.rows() > 0);
    assert(matches.ids.cols() > 0);
    assert(matches.ids.cols() == requestedPts.features.cols()); //nbpts
    assert(outlierWeights.rows() == matches.ids.rows());  // knn
    
    const int knn = outlierWeights.rows();
    const int dimFeat = requestedPts.features.rows();
    const int dimReqDesc = requestedPts.descriptors.rows();

    // Count points with no weights
    const int pointsCount = (outlierWeights.cwise() != 0.0).count();
    if (pointsCount == 0)
        throw ConvergenceError("ErrorMnimizer: no point to minimize");

    Matrix keptFeat(dimFeat, pointsCount);
    
    Matrix keptDesc;
    if(dimReqDesc > 0)
        keptDesc = Matrix(dimReqDesc, pointsCount);

    Matches keptMatches (Dists(1,pointsCount), Ids(1, pointsCount));
    OutlierWeights keptWeights(1, pointsCount);

    int j = 0;
    weightedPointUsedRatio = 0;
    for(int k = 0; k < knn; k++) // knn
    {
        for (int i = 0; i < requestedPts.features.cols(); ++i) //nb pts
        {
            if (outlierWeights(k,i) != 0.0)
            {
                if(dimReqDesc > 0)
                    keptDesc.col(j) = requestedPts.descriptors.col(i);
                
                keptFeat.col(j) = requestedPts.features.col(i);
                keptMatches.ids(0, j) = matches.ids(k, i);
                keptMatches.dists(0, j) = matches.dists(k, i);
                keptWeights(0,j) = outlierWeights(k,i);
                ++j;
                weightedPointUsedRatio += outlierWeights(k,i);
            }
        }
    }

    assert(j == pointsCount);

    this->pointUsedRatio = double(j)/double(knn*requestedPts.features.cols());
    this->weightedPointUsedRatio /= double(knn*requestedPts.features.cols());
    
    assert(dimFeat == sourcePts.features.rows());
    const int dimSourDesc = sourcePts.descriptors.rows();
    
    Matrix associatedFeat(dimFeat, pointsCount);
    Matrix associatedDesc;
    if(dimSourDesc > 0)
        associatedDesc = Matrix(dimSourDesc, pointsCount);

    // Fetch matched points
    for (int i = 0; i < pointsCount; ++i)
    {
        const int refIndex(keptMatches.ids(i));
        associatedFeat.col(i) = sourcePts.features.block(0, refIndex, dimFeat, 1);
        
        if(dimSourDesc > 0)
            associatedDesc.col(i) = sourcePts.descriptors.block(0, refIndex, dimSourDesc, 1);
    }

    lastErrorElements.reading = DataPoints(
        keptFeat, 
        requestedPts.featureLabels,
        keptDesc,
        requestedPts.descriptorLabels
    );
    lastErrorElements.reference = DataPoints(
        associatedFeat,
        sourcePts.featureLabels,
        associatedDesc,
        sourcePts.descriptorLabels
    );
    lastErrorElements.weights = keptWeights;
    lastErrorElements.matches = keptMatches;
    return lastErrorElements;
}

template struct PointMatcher<float>::ErrorMinimizer;
template struct PointMatcher<double>::ErrorMinimizer;