Monocular 3D Object Detection On Google

AP at 10' Elevation error

AP at 15' Azimuth error

Average Precision at 0.5 3D IoU

MPE

评测结果

各个模型在此基准测试上的表现结果

					Paper Title	Repository
EfficientNetLite + keypoint regressor	0.8584	0.7844	0.6512	0.0467	Objectron: A Large Scale Dataset of Object-Centric Videos in the Wild with Pose Annotations
Lin2021	0.8560	0.8398	0.7218	0.0520	-	-
MobilePose	0.6658	0.5088	0.4624	0.1001	MobilePose: Real-Time Pose Estimation for Unseen Objects with Weak Shape Supervision	-

0 of 3 row(s) selected.