4.6 Article

Stereo CenterNet-based 3D object detection for autonomous driving

期刊

NEUROCOMPUTING
卷 471, 期 -, 页码 219-229

出版社

ELSEVIER
DOI: 10.1016/j.neucom.2021.11.048

关键词

3D object detection; Stereo imagery; Photometric alignment

资金

  1. National Natural Science Foundation of China [61772068]
  2. Finance science and technology project of Hainan province [ZDYF2019009]
  3. China Postdoctoral Science Foundation [2020M680352]
  4. Guangdong Basic and Applied Basic Research Foundation [2020A1515110463]
  5. Scientific and Technological Innovation Foundation of Shunde Graduate School, USTB [2020BH011]
  6. USTB MatCom of Beijing Advanced Innovation Center for Materials Genome Engineering

向作者/读者索取更多资源

In this study, a 3D object detection method called Stereo CenterNet (SC) is proposed, which utilizes geometric information from stereo imagery. SC predicts the key points and utilizes 2D and 3D information to restore the object's bounding box in 3D space. Experimental results show that SC achieves the best trade-off between speed and accuracy.
Recently, three-dimensional (3D) detection based on stereo images has progressed remarkably; however, most advanced methods adopt anchor-based two-dimensional (2D) detection or depth estimation to address this problem. Nevertheless, high computational cost inhibits these methods from achieving real-time performance. In this study, we propose a 3D object detection method, Stereo CenterNet (SC), using geometric information in stereo imagery. SC predicts the four semantic key points of the 3D bound-ing box of the object in space and utilizes 2D left and right boxes, 3D dimension, orientation, and key points to restore the bounding box of the object in the 3D space. Subsequently, we adopt an improved photometric alignment module to further optimize the position of the 3D bounding box. Experiments conducted on the KITTI dataset indicate that the proposed SC exhibits the best speed-accuracy trade -off among advanced methods without using extra data. (c) 2021 Elsevier B.V. All rights reserved.

作者

我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。

评论

主要评分

4.6
评分不足

次要评分

新颖性
-
重要性
-
科学严谨性
-
评价这篇论文

推荐

暂无数据
暂无数据