Is there a list of all possible AWS Rekognition “scene” labels?

Question

Problem:

I want to separate out "objects" from "scenes" in AWS Rekognition API responses. Currently, the JSON responses lump &q