void TLD::selectObject(const Mat &img, Rect *bb)
{
//Delete old object
detectorCascade->release();
detectorCascade->objWidth = bb->width;
detectorCascade->objHeight = bb->height;
//Init detector cascade
detectorCascade->init();
currImg = img;
if(currBB)
{
delete currBB;
currBB = NULL;
}
currBB = tldCopyRect(bb);
currConf = 1;
valid = true;
initialLearning();
}
void TLD::selectObject(const Mat &img, Rect *bb)
{
//Delete old object
detectorCascade->release();
detectorCascade->objWidth = bb->width;
detectorCascade->objHeight = bb->height;
//Init detector cascade
detectorCascade->init();
#ifdef USE_HTLD
if(hTLDMaster->getNumOfObjects() > 0)
hTLDMaster->destroyTLDObject(tldObjId);
FAST_DET_INIT_PARAMS* fdi = detectorCascade->createFastDetInitStr();
FAST_TR_INIT_PARAMS* fti = medianFlowTracker->createFastTrInitStr();
tldObjId = hTLDMaster->createTLDObject(fti, fdi);
free((void*)fdi);
#endif
currImg = img;
if(currBB)
{
delete currBB;
currBB = NULL;
}
currBB = tldCopyRect(bb);
currConf = 1;
valid = true;
initialLearning();
}
void TLD::fuseHypotheses() {
Rect* trackerBB = medianFlowTracker->trackerBB;
int numClusters = detectorCascade->detectionResult->numClusters;
Rect* detectorBB = detectorCascade->detectionResult->detectorBB;
currBB = NULL;
currConf = 0;
valid = false;
float confDetector = 0;
if(numClusters == 1) {
confDetector = nnClassifier->classifyBB(currImg, detectorBB);
}
if(trackerBB != NULL) {
float confTracker = nnClassifier->classifyBB(currImg, trackerBB);
if(numClusters == 1 && confDetector > confTracker && tldOverlapRectRect(*trackerBB, *detectorBB) < 0.5) {
currBB = tldCopyRect(detectorBB);
currConf = confDetector;
} else {
currBB = tldCopyRect(trackerBB);
currConf = confTracker;
if(confTracker > nnClassifier->thetaTP) {
valid = true;
} else if(wasValid && confTracker > nnClassifier->thetaFP) {
valid = true;
}
}
} else if(numClusters == 1) {
currBB = tldCopyRect(detectorBB);
currConf = confDetector;
}
/*
float var = CalculateVariance(patch.values, nn->patch_size*nn->patch_size);
if(var < min_var) { //TODO: Think about incorporating this
printf("%f, %f: Variance too low \n", var, classifier->min_var);
valid = 0;
}*/
}
void TLD::fuseHypotheses(const Mat& colorImg) {
int numClusters = detectorCascade->detectionResult->numClusters;
Rect* detectorBB = detectorCascade->detectionResult->detectorBB;
deleteCurrentBB();
currConf = 0;
valid = false;
if (!detectorEnabled) {
if (!isTrackerValid)
{ return; }
currBB = tldCopyRect(&trackerBB);
valid = true;
return;
}
float confDetector = 0;
if (numClusters == 1)
{ confDetector = nnClassifier->classifyBB(currImg, detectorBB); }
if (isTrackerValid) {
float confTracker = nnClassifier->classifyBB(currImg, &trackerBB);
currBB = tldCopyRect(&trackerBB);
valid = true;
currConf = confTracker;
} else if (numClusters == 1) {
// reinit tracker
if (trackerEnabled) {
if (tracker->updateAt(colorImg, *detectorBB)) {
currConf = nnClassifier->classifyBB(currImg, detectorBB);
currBB = tldCopyRect(detectorBB);
valid = true;
runTracker = true;
} else {
deleteCurrentBB();
currConf = 0;
valid = false;
}
}
}
}
void TLD::selectObject(const Mat& img, Rect* bb) {
if (!tracker)
{ return; }
if (!rng) {
std::random_device rd;
std::shared_ptr<std::mt19937> g(new std::mt19937(rd()));
g->seed((unsigned long)seed);
rng = g;
}
rng->seed((unsigned long)seed);
Mat grayFrame;
cvtColor(img, grayFrame, CV_BGR2GRAY);
// initialize the image dimensions once
if (!hasImageDimensions) {
detectorCascade->imgWidth = grayFrame.cols;
detectorCascade->imgHeight = grayFrame.rows;
detectorCascade->imgWidthStep = static_cast<int>(grayFrame.step);
}
runTracker = true;
//Delete old object
detectorCascade->release();
detectorCascade->objWidth = bb->width;
detectorCascade->objHeight = bb->height;
if (bb->height < detectorCascade->minSize)
{ detectorCascade->minSize = bb->height; }
if (bb->width < detectorCascade->minSize)
{ detectorCascade->minSize = bb->width; }
//Init detector cascade
detectorCascade->init(rng);
tracker->reinit(img, *bb);
currImg = grayFrame;
deleteCurrentBB();
currBB = tldCopyRect(bb);
currConf = 1;
valid = true;
initialLearning();
}
