Abstract: The 3D Average Precision (3DAP) relies on the intersection over union between predictions and ground truth objects. However, camera-only detectors have limited depth accuracy, which may ...