In video annotation I’d like to have some keypoints related to a bounding box label - like skeletal keypoints related to a bounding box for the body. Is there a best practice for doing this in labelbox?
Thanks for the feedback, @john1. It’s in our roadmap. We will likely add this during q1 2021. Meanwhile, the best way to achieve this is by one of the two ways:
- Use nested classifications as a means to tie all objects together.
- Use post-processing to identify objects inside a bbox. (does not work with overlapping objects)