extract 3D coordinates of keypoints

Software Python
1

I’m starting to work with the ZED X camera and I’d like to know if it’s possible to extract the 3D coordinates of a keypoint using a custom YOLO detection model.

My custom YOLO model is already trained, tested, and validated to identify elements and keypoints. I can successfully obtain the 2D bounding boxes from detections with ZED, but I haven’t figured out how to retrieve the corresponding 3D coordinates.

I’ve tried using pyzed.sl.BodyData and pyzed.sl.CustomBoxObjectData, but I wasn’t able to make it work for my case.

I just need to figure out how to obtain the corresponding 3D keypoint coordinates.

2

Hi,

It’s currently not possible.

The custom detector feature of the ZEDK SDK currently only support Bounding boxes, not skeleton data.

3

Ok, thank you. I’m converting keypoints to bbox structure and it works for now.