An Empirical Analysis of Range for 3D Object Detection

Abstract

LiDAR-based 3D detection plays a vital role in autonomous navigation. Surprisingly, although autonomous vehicles (AVs) must detect both near-field objects (for collision avoidance) and far-field objects (for longer-term planning), contemporary benchmarks focus only on near-field 3D detection. However, AVs must detect far-field objects for safe navigation. In this paper, we present an empirical analysis of far-field 3D detection using the long-range detection dataset Argoverse 2.0 to better understand the problem, and share the following insight: near-field LiDAR measurements are dense and optimally encoded by small voxels, while far-field measurements are sparse and are better encoded with large voxels. We exploit this observation to build a collection of range experts tuned for near-vs-far field detection, and propose simple techniques to efficiently ensemble models for long-range detection that improve efficiency by 33% and boost accuracy by 3.2% CDS.

Cite

Text

Peri et al. "An Empirical Analysis of Range for 3D Object Detection." IEEE/CVF International Conference on Computer Vision Workshops, 2023. doi:10.1109/ICCVW60793.2023.00440

Markdown

[Peri et al. "An Empirical Analysis of Range for 3D Object Detection." IEEE/CVF International Conference on Computer Vision Workshops, 2023.](https://mlanthology.org/iccvw/2023/peri2023iccvw-empirical/) doi:10.1109/ICCVW60793.2023.00440

BibTeX

@inproceedings{peri2023iccvw-empirical,
  title     = {{An Empirical Analysis of Range for 3D Object Detection}},
  author    = {Peri, Neehar and Li, Mengtian and Wilson, Benjamin and Wang, Yu-Xiong and Hays, James and Ramanan, Deva},
  booktitle = {IEEE/CVF International Conference on Computer Vision Workshops},
  year      = {2023},
  pages     = {4076-4085},
  doi       = {10.1109/ICCVW60793.2023.00440},
  url       = {https://mlanthology.org/iccvw/2023/peri2023iccvw-empirical/}
}