Has anyone on here experience with #BoofCV?

I want to detect shapes based on png/svg "masks" (black image, transparent bg) but I can only find examples where the shape to detect is defined in code using some sort of coordinates.

Boosts are highly appreciated!