NICE_VisLearning/features/fpfeatures/ConvolutionFeature.cpp

/**
* @file ConvolutionFeature.cpp
* @brief convolutional feature
* @author Sven Sickert
* @date 10/13/2008

*/
#include <iostream>
#include <time.h>

#include "ConvolutionFeature.h"
#include "vislearning/cbaselib/FeaturePool.h"

using namespace OBJREC;

using namespace NICE;

/* Convolutional feature consists of shift parameter params[0] and the
   convolutional mask, which is stored in the rest of the parameter vector
   params */


/** simple constructor */
ConvolutionFeature::ConvolutionFeature ( )
{
    window_size_x = 15;
    window_size_y = 15;
    isColor = false;
    useSpatialPriors = false;

    initializeParameterVector();
}

/** alternative constructor */
ConvolutionFeature::ConvolutionFeature (
        const int wsize_x,
        const int wsize_y,
        const bool color,
        const bool prior )
{
    window_size_x = wsize_x;
    window_size_y = wsize_y;
    isColor = color;
    useSpatialPriors = prior;

    initializeParameterVector();
}

/** default constructor */
ConvolutionFeature::ConvolutionFeature ( const Config *conf )
{
    std::string section = "ConvolutionFeature";
    window_size_x = conf->gI ( section, "window_size_x", 15 );
    window_size_y = conf->gI ( section, "window_size_y", 15 );
    isColor = conf->gB ( section, "is_color", false );
    useSpatialPriors = conf->gB ( section, "use_spatial_priors", false );

    initializeParameterVector();
}

/** copy constructor */
ConvolutionFeature::ConvolutionFeature ( const ConvolutionFeature *confFeat )
{
    window_size_x = confFeat->window_size_x;
    window_size_y = confFeat->window_size_y;
    paramsLength = confFeat->paramsLength;
    isColor = confFeat->isColor;
    useSpatialPriors = confFeat->useSpatialPriors;
    numChannels = confFeat->numChannels;
    params = new NICE::Vector( paramsLength, 0.0 );

    int i = 0;
    for ( NICE::Vector::iterator it = confFeat->params->begin();
          it != confFeat->params->end(); ++it, i++ )
    {
        params[i] = *it;
    }
}

/** simple destructor */
ConvolutionFeature::~ConvolutionFeature ( )
{
    if ( params != NULL)
        delete params;
}


/** (re)initialize parameter vector */
void ConvolutionFeature::initializeParameterVector()
{
    if (window_size_x > 0 && window_size_y > 0)
    {
        if (isColor)
            numChannels = 3;
        else
            numChannels = 1;

        paramsLength = numChannels*window_size_x*window_size_y + 1;

        if (useSpatialPriors) paramsLength += 2;

        params = new NICE::Vector( paramsLength, (1.0/(double)(paramsLength-1) ) );
        params[0] = 1;
    }
    else
        std::cerr << "ConvolutionFeature::initializeVector: Size of window is Zero! Could not initialize..."
                  << std::endl;
}

bool ConvolutionFeature::isColorMode() const
{
    return isColor;
}

/** return parameter vector */
NICE::Vector ConvolutionFeature::getParameterVector() const
{
    NICE::Vector res = (*this->params);
    return res;
}

/** return feature vector */
NICE::Vector ConvolutionFeature::getFeatureVector( const Example *example ) const
{
    NICE::Vector vec(paramsLength, 1.0);

    NICE::MultiChannelImageT<double> * imgD = NULL;
    imgD = & example->ce->getDChannel( CachedExample::D_EOH );

    int xsize, ysize;
    example->ce->getImageSize( xsize, ysize );

    const int x = example->x;
    const int y = example->y;
    const int halfwsx = std::floor ( window_size_x / 2 );
    const int halfwsy = std::floor ( window_size_y / 2 );
    const int step = window_size_x*window_size_y;

    int k = 1;
    for ( int v = -halfwsy; v <= halfwsy; v++ )
        for ( int u = -halfwsx; u <= halfwsx; u++, k++ )
        {
            int uu = u;
            int vv = v;
            if (x+u < 0 || x+u >= xsize) uu=-u;
            if (y+v < 0 || y+v >= ysize) vv=-v;

            if ( x+uu > 0
                 && x+uu < xsize
                 && y+vv > 0
                 && y+vv < ysize
                 && k < vec.size() )
            {
                for ( int c = 0; c < numChannels; c++)
                    vec[k+c*step] = imgD->get(x+uu,y+vv,c);
            }
        }

    if (useSpatialPriors)
    {
        vec[paramsLength-2] = (double)x/(double)xsize;
        vec[paramsLength-1] = (double)y/(double)ysize;
    }

    return vec;
}

/** return length of parameter vector */
int ConvolutionFeature::getParameterLength() const
{
    return paramsLength;
}

void ConvolutionFeature::setRandomParameterVector ( )
{
    srand (time(NULL));
    for ( NICE::Vector::iterator it = params->begin();
          it != params->end(); ++it )
    {
        double b = (double) rand() / (double) RAND_MAX;
        *it = b;
    }
    params->normalizeL2();
}

/** set parameter vector */
void ConvolutionFeature::setParameterVector( const Vector & vec )
{
    if ( params->size() == vec.size() )
    {
        int i = 0;
        for ( NICE::Vector::iterator it = params->begin();
              it != params->end(); ++it, i++ )
        {
            *it = vec[i];
        }
        params->normalizeL2();
    }
    else
        std::cerr << "ConvolutionFeature::setParameterVector: Vector sizes do not match!"
                  << " expected: " << params->size() << ", got: " << vec.size()
                  << std::endl;

}

/** return feature value */
double ConvolutionFeature::val ( const Example *example ) const
{
    double val1 = 0.0;

    // is parameter vector and image data available?
    if (params == NULL)
    {
        std::cerr << "ConvolutionalFeature::val: Missing parameter vector!"
                  << std::endl;

        return val1;
    }

    NICE::Vector featVec = getFeatureVector ( example );

//    for ( int i = 0; i < featVec.size(); i++ )
//        val1 += featVec[i] * params->operator [](i);
    val1 = params->scalarProduct ( featVec );

    return val1;
}

/** creature feature pool */
void ConvolutionFeature::explode ( FeaturePool &featurePool, bool variableWindow ) const
{
    ConvolutionFeature *f = new ConvolutionFeature (
                this->window_size_x,
                this->window_size_y,
                this->isColor,
                this->useSpatialPriors );

    featurePool.addFeature(f);
}

/** clone current feature */
Feature *ConvolutionFeature::clone ( ) const
{
    ConvolutionFeature *f = new ConvolutionFeature (
                this->window_size_x,
                this->window_size_y,
                this->isColor,
                this->useSpatialPriors );

    f->setParameterVector( *params );

    return f;
}

Feature *ConvolutionFeature::generateFirstParameter () const
{
    return clone();
}

void ConvolutionFeature::restore ( std::istream & is, int format )
{
    is >> window_size_x;
    is >> window_size_y;
    is >> paramsLength;

    isColor = false;
    useSpatialPriors = false;
    numChannels = 1;

    if ( paramsLength == (window_size_x*window_size_y+3) )
    {
        useSpatialPriors = true;
    }
    else if ( paramsLength == (3*window_size_x*window_size_y+1) )
    {
        isColor = true;
        numChannels = 3;
    }
    else if ( paramsLength == (3*window_size_x*window_size_y+3) )
    {
        isColor = true;
        numChannels = 3;
        useSpatialPriors = true;
    }

    params = new NICE::Vector( paramsLength, 1.0 );
    for ( NICE::Vector::iterator it = params->begin();
          it != params->end(); ++it )
        is >> *it;
}

void ConvolutionFeature::store ( std::ostream & os, int format ) const
{
    os << "ConvolutionFeature "
       << window_size_x << " "
       << window_size_y << " "
       << paramsLength;

    for ( NICE::Vector::const_iterator it = params->begin();
          it != params->end(); ++it )
        os << ' ' << *it;

}

void ConvolutionFeature::clear ()
{
    params->clear();
}