search

ispc-lab / HRegNet

[ICCV 2021] HRegNet: A Hierarchical Network for Large-scale Outdoor LiDAR Point Cloud Registration
MIT License
90 stars 12 forks source link

How to get the rotation error? #4

Closed xiyue-wx closed 2 years ago

xiyue-wx commented 2 years ago

Thank for your generous sharing. However after I reproduced your experiment, I found that the results of the rotation error in the KITTI_SUMMARY is not consistent with the results in the paper. How did you get it from the results in the KITTI_SUMMARY?

FanLu97 commented 2 years ago

Hello, @xiyue-wx , Thanks for your interest in our work. The Rotation mean and Rotation std is the rotation error in the paper, it should be consistent with the paper. Could you please provide the reproduced results in your "kitti_summary" file?

xiyue-wx commented 2 years ago

The results in the summary are as follows： Dataset: kitti Translation threshold: 2.00 Rotation threshold: 5.00 Translation mean: 0.1237 Translation std: 0.1342 Rotation mean: 0.9911 Rotation std: 1.0370 Runtime: 107276.5362 Success rate: 0.9965

FanLu97 commented 2 years ago

It's weird. It seems that the translation error and success rate are consistent with my results. I will re-run the code and put the results here. Besides, I have updated the code today and maybe you can download the new version and re-run the experiments.

xiyue-wx commented 2 years ago

thanks, I will download the updated code and re-run the experiments

FanLu97 commented 2 years ago

I re-run the experiments with the current version and the results on KITTI dataset are listed below, which is slightly better than the version in the paper. Dataset: kitti Translation threshold: 2.00 Rotation threshold: 5.00 Translation mean: 0.1106 Translation std: 0.1123 Rotation mean: 0.2807 Rotation std: 0.2380 Runtime: 111652.4726 Success rate: 0.9971 Please let me know if you have any new results, thanks.

xiyue-wx commented 2 years ago

I re-run the updated experiments, it's weirder. Dataset: kitti Translation threshold: 2.00 Rotation threshold: 5.00 Translation mean: 0.0539 Translation std: 0.0699 Rotation mean: 0.9203 Rotation std: 1.0351 Runtime: 118950.0126 Success rate: 0.9988

FanLu97 commented 2 years ago

It's weird. Maybe you can e-mail me your contact information for more convenient communications. My email: lufan@tongji.edu.cn

FanLu97 commented 2 years ago

Close due to inactivity.

jiangchaokang commented 1 year ago

I re-run the updated experiments, it's weirder. Dataset: kitti Translation threshold: 2.00 Rotation threshold: 5.00 Translation mean: 0.0539 Translation std: 0.0699 Rotation mean: 0.9203 Rotation std: 1.0351 Runtime: 118950.0126 Success rate: 0.9988

I'm having the same problem as you. Have you solved this issue? I would like to know what the solution is? Looking forward to your reply！