Table 9

Performance comparison with state-of-the-art models on the Nuscenes val set for all 10 classes and all the 3D detection results are evaluated with average precision.

ModelsCarTruckCVBusTrailerBarrierMotor.Cyc.Ped.Cone
Voxel-RCNN, [6]68.8828.036.0454.3417.9925.3722.397.2159.5525.39
Ours69.2528.956.8354.3118.0426.5723.758.4659.5625.08

Note: The above results are reproduced by the publicly release model ([28]). Note that CV stands for construction vehicle and Cone for traffic cone.

or Create an Account

Close subscription notice
Close access options