Closed ahsandar closed 10 years ago
The HOGDescriptor getting wrapped is the one from opencv2/objedetect which has this strange behaviour.
Unfortunately, there no build in conversion to convert from std::vector
vec = std::Vector.new(FLOAT)
mat = Cv::Mat.new(vec.size,1,Cv::CV_32FC1,vec.data,Cv::Mat::AUTO_STEP)
Thanks , but can you add an example of Hog descriptor to the Git will be really helpful, cheers
Can we have an example of how to use HogDescriptor in ropencv, as it seems in opencv hog.setSVMDetector takes HOGDescriptor::getDefaultPeopleDetector() as an argument
while in ropencv the wrappers for HOGDescriptor::getDefaultPeopleDetector() returns Std::Vector::Float while hog.setSVMDetector expects Cv::Mat