Question on 3D Object Detection Models available in the Zed-Ros-Wrapper

Software ROS1
1

Dear StereoLabs,
I’m curious if there is any information about the Detection Algorithms used for the ROS-Wrapper (MULTI_CLASS_BOX_FAST, _MEDIUM, and _ACCURATE). From this Post I’m aware that the information I’m after was proprietary, but is this still the case?

I would be glad for any kind of information about the algorithm architecture, but mostly am interested in which deep learning model is used.

Best
Janosch

2

Hi @Jan95
Welcome to the Stereolabs community.
The ROS Wrapper uses the ZED SDK to obtain object detection information, so the information that you are searching for is not public.

3

That’s unfortunate, but thanks for the fast reply :slight_smile: