I'm trying to find a person sitting down so I need to train my own HOG and SVM. After reading how HOG and SVM's work I really can't find a simple implementation for creating my own custom HOG and SVM. I get that I need to crop positive and negative sample images to a specific size and places them in respective folders. This program (https://github.com/opencv/opencv/blob/master/samples/cpp/train_HOG.cpp) train_HOG.cpp, from OpenCV, looks like a clean implementation but I'm not sure where to start.
Can I install OpenCV on my Windows machine and compile it with Visual Studio? After it runs is the output a .XML file?
I'm using OpenCV and Python on a Raspberry Pi. I've been able to use other trained .XML files by placing them in the Data folder where the default people detector is.