OpenCV 3 HOG 检测信心？

Question

I am using the OpenCV 3 HOG people detector to detect people moving in front of my laptops web cam. The detection part works sort of fine but I would like to obtain the confidence from the HOG classifier, which I think should be possible.

I use the following code to obtain the bounding boxes for the detected objects:

std::vector< cv::Rect> found_locations_rect;
d_hog->detectMultiScale(rGpuImg, found_locations_rect);

According to the Intellisense tip it should be possible to extract the confidence using:

void detectMultiScale(cv::InputArray img, std::vector<cv::Rect> &found_locations, std::vector<double> *confidence = (std::vector<double> *)0);

But I am not sure how to declare and initialize the *confidence variable, can you please help me?

Answer 1

You can take a look at official OpenCV documentation , it declares this overload of detectMultiScale function ( CPU implementation):

virtual void cv::HOGDescriptor::detectMultiScale    (   InputArray  img,
std::vector< Rect > &   foundLocations,
std::vector< double > &     foundWeights,
double  hitThreshold = 0,
Size    winStride = Size(),
Size    padding = Size(),
double  scale = 1.05,
double  finalThreshold = 2.0,
bool    useMeanshiftGrouping = false 
)       const

and for GPU

virtual void cv::cuda::HOG::detectMultiScale    (   InputArray  img,
std::vector< Rect > &   found_locations,
std::vector< double > *     confidences = NULL 
)   

So you can simply call it ( CPU mode):

std::vector< cv::Rect> found_locations_rect;
std::vector<double> found_weights;
d_hog->detectMultiScale(mat, found_locations_rect, found_weights);

or ( GPU implementation):

std::vector< cv::Rect> found_locations_rect;
std::vector<double> confidences;
d_hog->detectMultiScale(rGpuImg, found_locations_rect, &confidences);

And if it will not work, OpenCV will throw an exception. You can display it like this:

try
{
    std::vector< cv::Rect> found_locations_rect;
    std::vector<double> confidences;
    d_hog->detectMultiScale(rGpuImg, found_locations_rect, &confidences);
}
catch(const std::exception& e)
{
    std::cout << e.what() << std::endl;
}

After that you can solve the problem