Yes it relies on the target being static when capturing the training data, but it’s ok for the background to move. We were actually surprised by how well it works on moving objects without being trained on them. In the post you can see Julius riding on a scooter and that is an unseen example with a detector that was only trained on static scooters.
Comments
This is only beneficial in static scenes right? Otherwise you can’t get free labels across the whole video.
Yes it relies on the target being static when capturing the training data, but it’s ok for the background to move. We were actually surprised by how well it works on moving objects without being trained on them. In the post you can see Julius riding on a scooter and that is an unseen example with a detector that was only trained on static scooters.