Hi,
some conversion rules:
- each interface is embedded inside a C++ template with "_di_" prefix. So, for example, TIEVisionObjectFind becomes _di_TIEVisionObjectFinder
- each method returns an HRESULT, so the actual return value is the last parameter
- the second-last parameter is always "false". This is an internal parameter.
- when a method expects a parent interface (like addClassifier that wants a TIEVisionBase, but we have a TIEVisionCascadeClassifier ), perform a cast, better a static_cast to avoid programming errors
Here is an example of conversion:
// check if "ievision.dll" is available
if (!IEVisionAvailable())
{
ShowMessage("ievision.dll not loaded!");
exit;
}
// load two face detectors
_di_TIEVisionObjectsFinder objectsFinder;
IEVisionLib->createObjectsFinder(false, objectsFinder);
_di_TIEVisionCascadeClassifier faceDetector1, faceDetector2;
IEVisionLib->createCascadeClassifier(":FRONTALFACEALTTREE", false, faceDetector1);
IEVisionLib->createCascadeClassifier(":FRONTALFACEDEFAULT", false, faceDetector2);
objectsFinder->addClassifier("face detector 1", static_cast<TIEVisionBase *>(faceDetector1));
objectsFinder->addClassifier("face detector 2", static_cast<TIEVisionBase *>(faceDetector2));
// detect objects
objectsFinder->findIn(ImageEnView1->IEBitmap->GetIEVisionImage());
// merge intersecting rectangles of all searched objects
_di_TIEVisionVectorRect rects;
objectsFinder->mergeAllRects(false, rects);
// loop among rectangles
int rectCount;
rects->size(false, rectCount);
for (int i=0; i != rectCount; ++i)
{
TIEVisionRect rectangle;
rects->getRect(i, false, rectangle);
// ... do something with the rectangle coordinates and size
ImageEnView1->IEBitmap->Canvas->Pen->Color = clRed;
ImageEnView1->IEBitmap->Canvas->Brush->Style = bsClear;
ImageEnView1->IEBitmap->Canvas->Rectangle(rectangle.x, rectangle.y, rectangle.x+rectangle.width, rectangle.y+rectangle.height);
}
Search
Face Detection C++
Posted - Mar 10 2012 : 08:09:49
