DiffuBox: Refining 3D Object Detection with Point Diffusion
Abstract
Ensuring robust 3D object detection and localization is crucial for many applications in robotics and autonomous driving. Recent models, however, face difficulties in maintaining high performance when applied to domains with differing sensor setups or geographic locations, often resulting in poor localization accuracy due to domain shift. To overcome this challenge, we introduce a novel diffusion-based box refinement approach. This method employs a domain-agnostic diffusion model, conditioned on the LiDAR points surrounding a coarse bounding box, to simultaneously refine the box's location, size, and orientation. We evaluate this approach under various domain adaptation settings, and our results reveal significant improvements across different datasets, object classes and detectors. Our PyTorch implementation is available at https://github.com/cxy1997/DiffuBox.
Cite
Text
Chen et al. "DiffuBox: Refining 3D Object Detection with Point Diffusion." Neural Information Processing Systems, 2024. doi:10.52202/079017-3293Markdown
[Chen et al. "DiffuBox: Refining 3D Object Detection with Point Diffusion." Neural Information Processing Systems, 2024.](https://mlanthology.org/neurips/2024/chen2024neurips-diffubox/) doi:10.52202/079017-3293BibTeX
@inproceedings{chen2024neurips-diffubox,
title = {{DiffuBox: Refining 3D Object Detection with Point Diffusion}},
author = {Chen, Xiangyu and Liu, Zhenzhen and Luo, Katie Z and Datta, Siddhartha and Polavaram, Adhitya and Wang, Yan and You, Yurong and Li, Boyi and Pavone, Marco and Chao, Wei-Lun and Campbell, Mark and Hariharan, Bharath and Weinberger, Kilian Q.},
booktitle = {Neural Information Processing Systems},
year = {2024},
doi = {10.52202/079017-3293},
url = {https://mlanthology.org/neurips/2024/chen2024neurips-diffubox/}
}