Dear NVIDIA Developers,
my question is: can the Human Pose Estimation Application output JSON files for every frame which contain the location of every key point or one JSON file which contains the location of every key point through the entire video? I’m refering to this pose estimation application.
So there’s two variants to this:
- a JSON file for every frame. In each of the JSON files there should be the locations of all key points (
- one large JSON file where locations of all key points through all the frames are
Is that supported?