I’m new to Python and machine learning and I have the following problem: I have annotated data in the COCO .json format. In this case, it is the surface area of corals on underwater photos that are alive and parts of corals that are dead. Sometimes, the masks overlap.

I’m new to Python and machine learning and I have the following problem: I have annotated data in the COCO .json format. In this case, it is the surface area of corals on underwater photos that are alive and parts of corals that are dead. Sometimes, the masks overlap. I want to subtract the area I annotated as “dead” from the areas I annotated as “alive”. On each photo, only one coral is annotated, sometimes I did this in multiple parts, therefore I also want to merge the masks of each class before the subtraction. The annotation class “dead” only occurs in a subset of the images. Could somebody point me in the right direction on how to do this?