0
$\begingroup$

I have a project where the data is annotated with key points without any bounding boxes. If I want only the key points without the bounding box around the object, is that possible or each object should have both bbox and key points. Can I train Yolov8 with key points detection without object detection (bounding boxes)?

If the answer is yes, how the annotation file should look like, please give me an example.

$\endgroup$

1 Answer 1

0
$\begingroup$

According to this github issue, it's not yet possible to train the model without any bounding box information unless you make changes to the architecture, and evaluation metrics. So the simple answer is "No, you can't train the model without box info".

$\endgroup$
1
  • $\begingroup$ that was me who asked the question :) $\endgroup$ Commented Jun 22, 2023 at 20:43

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .