MODE:3D Moving Object Detection with Event LiDAR Points
This is the official implementation of:
MODE:3D Moving Object Detection with Event LiDAR Points
Visualization
Main results
KITTI 3D MOT(dynamic objects)
| BEV | Car | Car | Car | Cyclist | Cyclist | Cyclist | Pedestrian | Pedestrian | Pedestrian |
|---|---|---|---|---|---|---|---|---|---|
| Difficulty | easy | moderate | hard | easy | moderate | hard | easy | moderate | hard |
| SECOND | 98.96 | 91.86 | 89.31 | 92.67 | 88.17 | 85.92 | 76.28 | 73.91 | 69.23 |
| PointPillars | 96.49 | 92.90 | 92.27 | 90.88 | 84.04 | 82.17 | 77.24 | 75.57 | 72.73 |
| Part-A2 | 96.06 | 93.51 | 91.06 | 91.13 | 86.04 | 85.21 | 74.75 | 71.84 | 67.42 |
Citation
Please consider citing our work as follows if it is helpful.
@inproceedings{available soon,
title=,
author={Yihang Li*, Yu Rong*, Fu Zhang},
booktitle={available soon},
year={2023}
}
Acknowledgments
This project is based on the following codebases.