search

Zengyi-Qin / MonoGRNet

MonoGRNet: A Geometric Reasoning Network for Monocular 3D Object Detection and Localization | KITTI
Apache License 2.0
245 stars 48 forks source link

The results in this paper are different from those in Kitti #33

Closed Xiaohuangya317 closed 4 years ago

Xiaohuangya317 commented 5 years ago

In the paper,when AP3D=0.7, rusults are 13.88 10.19 7.62, running time is 0.06s. But on KITTI website rusults are 9.61 5.74 4.25, running time is 0.04s.

Zengyi-Qin commented 5 years ago

On KITTI website it wrotes "Note 2: On 08.10.2019, we have followed the suggestions of the Mapillary team in their paper Disentangling Monocular 3D Object Detection and use 40 recall positions instead of the 11 recall positions proposed in the original Pascal VOC benchmark. This results in a more fair comparison of the results, please check their paper. The last leaderboards right before this change can be found here: Object Detection Evaluation, 3D Object Detection Evaluation, Bird's Eye View Evaluation."

Zengyi-Qin commented 5 years ago

The numbers in our paper are reported before the change. And it's reported on the val set instead of the test set. Regarding the inference time, you could find that the GPU we used are different. On the website we used Nvidia P40 that is more powerful

KunikaValecha commented 4 years ago

How can we test on a dataset other than given in this repo?