This video shows you how to quickly annotate objects in a video using CVAT’s skeleton tool.

‍

The process involves:

- creating a new task and label name,

- highlighting features on the object in a specific frame (e.g. hand, elbow), and then

- iterating over every few frames to correct for the object’s movement.

CVAT then uses interpolation to annotate the “in-between” frames, saving the user time.

We are waiting for your feedback here:

Discord

LinkedIn

Gitter

GitHub

‍

You can find more information at our YouTube channel