DJNing / See_beyond_seeing

[ICRA 2024] Robust 3D Object Detection from LiDAR-Radar Point Clouds Via Cross-Modal Feature Augmentation
18 stars 0 forks source link

LiDAR results under the IoU of 0.5, 0.25, 0.25 #2

Open XiangyPeng opened 1 month ago

XiangyPeng commented 1 month ago

Hi, thank you for your great work! I noticed that your results of LiDAR are based on the IoU of 0.7, 0.5, and 0.5, as Kitti format. May I kindly ask if you have the results under the IoU of 0.5, 0.25, and 0.25 for cars, cyclists, and pedestrians for LiDAR point clouds? If not, could you please provide the configuration file or some guidance to obtain the results of 0.5, 0.25, and 0.25 IoU for LiDAR data?

DJNing commented 1 month ago

Hi, thanks for your interests in my work.

One simple way to obtain the required metric would be to modify the threshold here: https://github.com/DJNing/See_beyond_seeing/blob/0764453887a0ea85e08fcf2f9d916f7ea31ef8c0/pcdet/datasets/vod/vod_object_eval_python/eval.py#L639