This needs updated badly.
I just started using opencv, and there is a large pitfall where training the classifier doesn't seem to have the right instructions attached.

https://docs.opencv.org/4.5.2/dc/d88/tutorial_traincascade.html

In the issue below, this is discussed. Note the date.
#13231

So I'm just seeking clarification as a beginner, should I not be using the object classifer in opencv ? Should I be using some alternative ? If the classifer is fine, what should I be doing to prepare positive samples ?