Skip to main content
All CollectionsTutorialsAnnotate Media
Annotate Full Motion Video
Annotate Full Motion Video

Create training data for object detection or segmentation models on video

T
Written by Taylor Maggos
Updated over a month ago

Within our platform, you can label/annotate still imagery and full motion video. Annotating full motion video is very similar to annotating imagery with one additional feature!

Whether you are annotating an object detection, segmentation or classification model, the base principals are the same -- drawing a bounding box around the object or creating a polygon. However, where it differs for video is you will then Copy and Paste those annotations to the rest of the frames in the video.

We have built in an interpolation tool which will allow you to make one label on frame 1, copy and paste it to frame 100 and the interpolation tool will automatically fill in all the frames in-between! Then you'll just have to go through and edit any of the labels to make sure they are capturing everything you need labeled.

Here is a short demo of how to annotate video and use the interpolation tool: How To Annotate Full Motion Video

Here is a snapshot of the annotation interface and the interpolation tool in action to give you an idea of how it works:

Did this answer your question?