中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 自动化研究所
  3. 中国科学院自动化研究所
  4. 学术期刊
  5. International Journal of Automation and Computing
On Robust Cross-view Consistency in Self-supervised Monocular Depth Estimation

文献类型：期刊论文

作者Haimei Zhao1; Jing Zhang1; Zhuo Chen3; Bo Yuan2; Dacheng Tao1
刊名Machine Intelligence Research
出版日期2024
卷号21期号:3页码:495-513
关键词3D vision, depth estimation, cross-view consistency, self-supervised learning, monocular perception
ISSN号2731-538X
DOI10.1007/s11633-023-1474-0
英文摘要Remarkable progress has been made in self-supervised monocular depth estimation (SS-MDE) by exploring cross-view consistency, e.g., photometric consistency and 3D point cloud consistency. However, they are very vulnerable to illumination variance, occlusions, texture-less regions, as well as moving objects, making them not robust enough to deal with various scenes. To address this challenge, we study two kinds of robust cross-view consistency in this paper. Firstly, the spatial offset field between adjacent frames is obtained by reconstructing the reference frame from its neighbors via deformable alignment, which is used to align the temporal depth features via a depth feature alignment (DFA) loss. Secondly, the 3D point clouds of each reference frame and its nearby frames are calculated and transformed into voxel space, where the point density in each voxel is calculated and aligned via a voxel density alignment (VDA) loss. In this way, we exploit the temporal coherence in both depth feature space and 3D voxel space for SS-MDE, shifting the “point-to-point” alignment paradigm to the “region-to-region” one. Compared with the photometric consistency loss as well as the rigid point cloud alignment loss, the proposed DFA and VDA losses are more robust owing to the strong representation power of deep features as well as the high tolerance of voxel density to the aforementioned challenges. Experimental results on several outdoor benchmarks show that our method outperforms current state-of-the-art techniques. Extensive ablation study and analysis validate the effectiveness of the proposed losses, especially in challenging scenes. The code and models are available at https://github.com/sunnyHelen/RCVC-depth.
源URL[http://ir.ia.ac.cn/handle/173211/56478]  
专题自动化研究所_学术期刊_International Journal of Automation and Computing
作者单位1.School of Computer Science, University of Sydney, Sydney 2008, Australia
2.School of Information Technology & Electrical Engineering, University of Queensland, Brisbane 4072, Australia
3.Shenzhen International Graduate School, Tsinghua University, Shenzhen 518055, China
推荐引用方式
GB/T 7714		 Haimei Zhao,Jing Zhang,Zhuo Chen,et al. On Robust Cross-view Consistency in Self-supervised Monocular Depth Estimation[J]. Machine Intelligence Research,2024,21(3):495-513.
APA Haimei Zhao,Jing Zhang,Zhuo Chen,Bo Yuan,&Dacheng Tao.(2024).On Robust Cross-view Consistency in Self-supervised Monocular Depth Estimation.Machine Intelligence Research,21(3),495-513.
MLA Haimei Zhao,et al."On Robust Cross-view Consistency in Self-supervised Monocular Depth Estimation".Machine Intelligence Research 21.3(2024):495-513.

入库方式： OAI收割

来源：自动化研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。