Template Matching

[ysnist]

I can say that it's a great plugin, but 99% of the users don't know how to model as you guessed. Instead of training a model, I think it would be great to match the image of an object cropped from an existing frame with the template matching available in opencv. You can do this as an additional feature

I believe it will be more user-friendly and I hope you will give standard users a chance.
For example, if I want to censor the cigarette packet in the image I watch on web cam, it should be very difficult to train a model for this