EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Learning Representative Features by Deep Attention Network for 3D Point Cloud Registration

Xiaokai Xia, Zhiqiang Fan, Gang Xiao, Fangyue Chen, Yu Liu, Yiheng Hu

Sensors · 20.04.2023 · 23 (8) · 4123

DOI: 10.3390/s23084123 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 20.04.2023. Verfasst von Xiaokai Xia und 5 weiteren Autor:innen. Beteiligte Einrichtung: Beijing Institute of System Engineering, Beijing 100101, China. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Three-dimensional point cloud registration, which aims to find the transformation that best aligns two point clouds, is a widely studied problem in computer vision with a wide spectrum of applications, such as underground mining. Many learning-based approaches have been developed and have demonstrated their effectiveness for point cloud registration. Particularly, attention-based models have achieved outstanding performance due to the extra contextual information captured by attention mechanisms. To avoid the high computation cost brought by attention mechanisms, an encoder–decoder framework is often employed to hierarchically extract the features where the attention module is only applied in the middle. This leads to the compromised effectiveness of the attention module. To tackle this issue, we propose a novel model with the attention layers embedded in both the encoder and decoder stages. In our model, the self-attentional layers are applied in the encoder to consider the relationship between points inside each point cloud, while the decoder utilizes cross-attentional layers to enrich features with contextual information. Extensive experiments conducted on public datasets prove that our model is able to achieve quality results on a registration task.

Schlagwörter

Stichwörter der Autor:innen

deep learningfeature extractionpoint cloudregistration

Autor:innen und Institutionen

  • Beijing Institute of System Engineering, Beijing 100101, China.
  • Artificial Intelligence Institute of China Electronics Technology Group Corporation, Beijing 100041, China.
  • State Key Laboratory of Software Development Environment, Beihang University, Beijing 100191, China.
  • School of Computer Science and Engineering, University of New South Wales, Sydney, NSW 2052, Australia.

Zitieren

APA

Xia, X., Fan, Z., Xiao, G., Chen, F., Liu, Y., & Hu, Y. (2023). Learning Representative Features by Deep Attention Network for 3D Point Cloud Registration. Sensors, 23(8), 4123. https://doi.org/10.3390/s23084123

Vancouver

Xia X, Fan Z, Xiao G, Chen F, Liu Y, Hu Y. Learning Representative Features by Deep Attention Network for 3D Point Cloud Registration. Sensors. 2023;23(8):4123. doi:10.3390/s23084123

BibTeX
@article{Xia2023,
  author = {Xia, Xiaokai and Fan, Zhiqiang and Xiao, Gang and Chen, Fangyue and Liu, Yu and Hu, Yiheng},
  title = {Learning Representative Features by Deep Attention Network for 3D Point Cloud Registration},
  journal = {Sensors},
  year = {2023},
  volume = {23},
  number = {8},
  pages = {4123},
  doi = {10.3390/s23084123}
}
Kostenlos starten →