react-image-annotate | Classify , tag images | Computer Vision library

I am working on a quiz tool where one of the questions is of image type.

Imagine you are creating the test for your students and you upload an image of a car. You then mark two doors of the car and your question is - Click on one of the doors of the car below. The student, when on this question, will see the same image without any markings/annotations. If the student clicks on the door he/she gets the point, not otherwise. Note: The car doors are random shapes and can be a polygon and preferably can be drawn using a brush tool, sort of.

Here are some thoughts I have but is there a better way or a library that suits this?:

The UI is built in React, the data format can be anything as long as it works.

1. A library that could draw and capture Polygons on an image like this or this. I can then run a "point lies inside polygon" algorithm like this. But that also sounds like overkill. (And I am not sure if they distinguish between multiple polygons cleanly).

2. A library that does both the tasks of option 1.

3. Anything else better?

Sorry if this is asked before, but my question is more React focused and I couldn't find an answer.