Skip to content

O
opencv_contrib
Commit b904b5b4 authored by jaco's avatar jaco
Browse files
Options

efficient structures access restored

parent e56f2870
Hide whitespace changes
Inline Side-by-side
Showing with 33 additions and 44 deletions
modules/saliency/src/motionSaliencyBinWangApr2014.cpp
View file @ b904b5b4
......@@ -61,10 +61,10 @@ MotionSaliencyBinWangApr2014::MotionSaliencyBinWangApr2014()
K = 3; // Number of background model template
N = 4; // NxN is the size of the block for downsampling in the lowlowResolutionDetection
alpha = 0.01; // Learning rate
L0 = 6000; // Upper-bound values for C0 (efficacy of the first template (matrices) of backgroundModel
L1 = 4000; // Upper-bound values for C1 (efficacy of the second template (matrices) of backgroundModel
L0 = 300; // Upper-bound values for C0 (efficacy of the first template (matrices) of backgroundModel
L1 = 200; // Upper-bound values for C1 (efficacy of the second template (matrices) of backgroundModel
thetaL = 50; // T0, T1 swap threshold
thetaA = 200;
thetaA = 50;
gamma = 3;
neighborhoodCheck = true;
......@@ -74,31 +74,26 @@ MotionSaliencyBinWangApr2014::MotionSaliencyBinWangApr2014()
bool MotionSaliencyBinWangApr2014::init()
{
epslonPixelsValue = Mat( imgSize->height, imgSize->width, CV_32F );
potentialBackground = Mat( imgSize->height, imgSize->width, CV_32FC2 );
//backgroundModel = std::vector<Mat>( K + 1, Mat::zeros( imgSize->height, imgSize->width, CV_32FC2 ) );
epslonPixelsValue = Mat( imgSize->height, imgSize->width, CV_32F, Scalar( 70 ) );
// Median of range [18, 80] advised in reference paper.
// Since data is even, the median is estimated using two values ​​that occupy
// the position (n / 2) and ((n / 2) +1) (choose their arithmetic mean).
//TODO set to nan
potentialBackground.setTo( 0 );
potentialBackground = Mat( imgSize->height, imgSize->width, CV_32FC2, Scalar( 0 ) );
//backgroundModel = std::vector<Mat>( K + 1, Mat::zeros( imgSize->height, imgSize->width, CV_32FC2 ) );
backgroundModel.resize(K+1);
backgroundModel.resize( K + 1 );
//TODO set to nan
for ( int i = 0; i < K + 1; i++ )
{
Mat* tmpm = new Mat;
tmpm->create(imgSize->height, imgSize->width, CV_32FC2);
tmpm->setTo(0);
tmpm->create( imgSize->height, imgSize->width, CV_32FC2 );
tmpm->setTo( 0 );
Ptr<Mat> tmp = Ptr<Mat>( tmpm );
backgroundModel[i] = tmp;
}
epslonPixelsValue.setTo( 70 ); // Median of range [18, 80] advised in reference paper.
// Since data is even, the median is estimated using two values ​​that occupy
// the position (n / 2) and ((n / 2) +1) (choose their arithmetic mean).
/* epslonPixelsValue = Mat::zeros( imgSize->height, imgSize->width, CV_8U );
potentialBackground = Mat::NAN( imgSize->height, imgSize->width, CV_32FC2 );
backgroundModel = std::vector<Mat>( 4, Mat::zeros( imgSize->height, imgSize->width, CV_32FC2 ) );*/
return true;
}
......@@ -142,8 +137,6 @@ bool MotionSaliencyBinWangApr2014::fullResolutionDetection( const Mat& image2, M
float* pEpslon;
float* pMask;
int countDec = 0;
// Scan all pixels of image
for ( int i = 0; i < image.rows; i++ )
{
......@@ -167,7 +160,7 @@ bool MotionSaliencyBinWangApr2014::fullResolutionDetection( const Mat& image2, M
int counter = 0;
for ( size_t z = 0; z < backgroundModel.size(); z++ )
{
counter += backgroundModel.at(z)->at<Vec2f>( i, j )[1];
counter += backgroundModel.at( z )->at<Vec2f>( i, j )[1];
}
if( counter != 0 ) //if at least the first template is activated / initialized
......@@ -176,40 +169,38 @@ bool MotionSaliencyBinWangApr2014::fullResolutionDetection( const Mat& image2, M
// scan background model vector
for ( size_t z = 0; z < backgroundModel.size(); z++ )
{
float currentB;
float currentC;
float* currentB;
float* currentC;
// TODO replace "at" with more efficient matrix access
currentB = (backgroundModel.at(z)->at<Vec2f>( i, j )[0]);
currentC = (backgroundModel.at(z)->at<Vec2f>( i, j )[1]);
currentB = &( backgroundModel.at( z )->at<Vec2f>( i, j )[0] );
currentC = &( backgroundModel.at( z )->at<Vec2f>( i, j )[1] );
if( i == 50 && j == 50 )
{
cout << "zeta:" << z << " currentB :" << currentB << endl << "currentC :" << currentC << endl;
cout << "zeta:" << z << " currentB :" << *currentB << endl << "currentC :" << *currentC << endl;
}
//continue;
if( currentC > 0 ) //The current template is active
if( *currentC > 0 ) //The current template is active
{
//cout<< "DIFFERENCE: "<<abs( currentPixelValue - ( *currentB ) )<<endl;
// If there is a match with a current background template
if( abs( currentPixelValue - ( currentB ) ) < currentEpslonValue && !backgFlag )
if( abs( currentPixelValue - ( *currentB ) ) < currentEpslonValue && !backgFlag )
{
// The correspondence pixel in the BF mask is set as background ( 0 value)
//highResBFMask.at<uchar>( i, j ) = 0;
pMask[j] = 0;
//if( ( *currentC < L0 && z == 0 ) || ( *currentC < L1 && z == 1 ) || ( z > 1 ) )
(backgroundModel.at(z)->at<Vec2f>( i, j )[1]) = (backgroundModel.at(z)->at<Vec2f>( i, j )[1]) + 1; // increment the efficacy of this template
if( ( *currentC < L0 && z == 0 ) || ( *currentC < L1 && z == 1 ) || ( z > 1 ) ){
*currentC += 1; // increment the efficacy of this template
}
(backgroundModel.at(z)->at<Vec2f>( i, j )[0]) = ( ( 1 - alpha ) * ( currentB ) ) + ( alpha * currentPixelValue ); // Update the template value
*currentB = ( ( 1 - alpha ) * ( *currentB ) ) + ( alpha * currentPixelValue ); // Update the template value
backgFlag = true;
//break;
}
else
{
if( z == 0 )
countDec++;
(backgroundModel.at(z)->at<Vec2f>( i, j )[1]) = (backgroundModel.at(z)->at<Vec2f>( i, j )[1]) - 1; // decrement the efficacy of this template
*currentC -= 1; // decrement the efficacy of this template
}
}
......@@ -217,8 +208,8 @@ bool MotionSaliencyBinWangApr2014::fullResolutionDetection( const Mat& image2, M
if( i == 50 && j == 50 )
{
cout << "DOPO IF: " << endl;
cout << "zeta:" << z << " currentB_A :" << &currentB << endl << "currentC_A :" << &currentC<<endl;
cout << "zeta:" << z << " currentB :" << (backgroundModel.at(z)->at<Vec2f>( i, j )[0]) << endl << "currentC :" << (backgroundModel.at(z)->at<Vec2f>( i, j )[1]) << endl<<endl;
cout << "zeta:" << z << " currentB_A :" << currentB << endl << "currentC_A :" << currentC << endl;
cout << "zeta:" << z << " currentB :" << *currentB << endl << "currentC :" << *currentC << endl << endl;
}
} // end "for" cicle of template vector
......@@ -281,7 +272,7 @@ bool MotionSaliencyBinWangApr2014::lowResolutionDetection( const Mat& image, Mat
for ( int z = 0; z < N_DS; z++ )
{
// Select the current template 2 channel matrix, select ROI and compute the mean for each channel separately
Mat roiTemplate = (*(backgroundModel[z]))( roi );
Mat roiTemplate = ( * ( backgroundModel[z] ) )( roi );
Scalar templateMean = mean( roiTemplate );
currentB = templateMean[0];
currentC = templateMean[1];
......@@ -413,8 +404,8 @@ bool MotionSaliencyBinWangApr2014::templateReplacement( const Mat& finalBFMask,
}
else
{
thetaA = 200;
neighborhoodCheck = false;
thetaA = 50;
neighborhoodCheck = true;
}
float roiSize = 3; // FIXED ROI SIZE, not change until you first appropriately adjust the following controls in the EVALUATION section!
......@@ -542,10 +533,8 @@ bool MotionSaliencyBinWangApr2014::templateReplacement( const Mat& finalBFMask,
}
else
{
((backgroundModel.at(backgroundModel.size() - 1))->at<Vec2f>( i, j ))[0] = potentialBackground.at<Vec2f>( i, j )[0];
((backgroundModel.at(backgroundModel.size() - 1))->at<Vec2f>( i, j ))[1] = potentialBackground.at<Vec2f>( i, j )[1];
potentialBackground.at<Vec2f>( i, j )[0] = 0;
potentialBackground.at<Vec2f>( i, j )[1] = 0;
backgroundModel[backgroundModel.size() - 1]->at<Vec2f>( i, j ) = potentialBackground.at<Vec2f>( i, j );
potentialBackground.at<Vec2f>( i, j ) = 0;
//((backgroundModel.at(0))->at<Vec2f>( i, j ))[1] = 3;
}
} // close if of EVALUATION
......
Markdown is supported
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