I’m using a ZED X with a custom object detector, and I’m ingesting the detections into the ZED SDK using the Object Detection API (custom bounding boxes). I can retrieve the distance of the detected objects, but I cannot find a way to access the confidence of the depth measurement associated with those objects.
Is there a way to directly obtain depth confidence from the ZED SDK for detected objects, without manually processing the depth map?
One possible workaround would be to retrieve the per-pixel depth confidence and compute the object distance and confidence manually within each bounding box. However for efficiency reasons I would prefer to avoid this approach.
Is there a recommended or built-in method to access depth confidence for object detection results?