So i'm really new to computer-vision and I think I know my answer but I wanted to ask first.
Im using detectron2 and I'm just getting started but is it possible to have one object belong to multiple classes?
For example: Lets say I was detecting attributes about people for some reason
So a person might have several classes. Lets say I want them to trigger 'person', 'male', and 'jim'
Using image segmentation it doesnt seem like the neural nets can handle it, (can object recognition vs the instance segmentation im looking for? I dont actually need the masks as much as detecting all the attributes of my objects)
right now I'm drawing polyginal bounding boxes and using a mask-rcnn model through detectron2 and i'm still just playing with it but that seems to be the case. If I train it to know jim, coming back and saying that this polygon is also a person gets no results.
I feel like what I'm going to have to do is run multiple models side by side on the same image to get all the results I want (but oof, then I'm going to have to figure out how to add all the mask areas together to derive all the classes an object can have)
edit: doing some more research and digging around I guess what Im trying to find is a model that can do multi-label image segmentation
