Depth Completion using Geometry-Aware Embedding

被引:21
|
作者
Du, Wenchao [1 ]
Chen, Hu [1 ]
Yang, Hongyu [1 ]
Zhang, Yi [1 ]
机构
[1] Sichuan Univ, Coll Comp Sci, Chengdu 610065, Peoples R China
关键词
NETWORK; SPARSE;
D O I
10.1109/ICRA46639.2022.9811556
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Exploiting internal spatial geometric constraints of sparse LiDARs is beneficial to depth completion, however, has been not explored well. This paper proposes an efficient method to learn geometry-aware embedding, which encodes the local and global geometric structure information from 3D points, e.g., scene layout, object's sizes and shapes, to guide dense depth estimation. Specifically, we utilize the dynamic graph representation to model generalized geometric relationship from irregular point clouds in a flexible and efficient manner. Further, we joint this embedding and corresponded RGB appearance information to infer missing depths of the scene with well structure-preserved details. The key to our method is to integrate implicit 3D geometric representation into a 2D learning architecture, which leads to a better trade-off between the performance and efficiency. Extensive experiments demonstrate that the proposed method outperforms previous works and could reconstruct fine depths with crisp boundaries in regions that are over-smoothed by them. The ablation study gives more insights into our method that could achieve significant gains with a simple design, while having better generalization capability and stability. The code is available at https://github.com/Wenchao-Du/GAENet.
引用
收藏
页码:8680 / 8686
页数:7
相关论文
共 50 条
  • [41] A Novel Framework for Learning Geometry-Aware Kernels
    Pan, Binbin
    Chen, Wen-Sheng
    Xu, Chen
    Chen, Bo
    IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2016, 27 (05) : 939 - 951
  • [42] Geometry-aware Compensation Scheme for Morphing Drones
    Fabris, Amedeo
    Kleber, Kevin
    Falanga, Davide
    Scaramuzza, Davide
    2021 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA 2021), 2021, : 592 - 598
  • [43] Adversarial Geometry-Aware Human Motion Prediction
    Gui, Liang-Yan
    Wang, Yu-Xiong
    Liang, Xiaodan
    Moura, Jose M. F.
    COMPUTER VISION - ECCV 2018, PT IV, 2018, 11208 : 823 - 842
  • [44] Geometry-Aware Learning of Maps for Camera Localization
    Brahmbhatt, Samarth
    Gu, Jinwei
    Kim, Kihwan
    Hays, James
    Kautz, Jan
    2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, : 2616 - 2625
  • [45] Geometry-Aware Cell Detection with Deep Learning
    Jiang, Hao
    Li, Sen
    Liu, Weihuang
    Zheng, Hongjin
    Liu, Jinghao
    Zhang, Yang
    MSYSTEMS, 2020, 5 (01)
  • [46] Disentangling Geometry and Appearance with Regularised Geometry-Aware Generative Adversarial Networks
    Tran, Linh
    Kossaifi, Jean
    Panagakis, Yannis
    Pantic, Maja
    INTERNATIONAL JOURNAL OF COMPUTER VISION, 2019, 127 (6-7) : 824 - 844
  • [47] Geometry-aware 3D pose transfer using transformer autoencoder
    Liu, Shanghuan
    Gai, Shaoyan
    Da, Feipeng
    Waris, Fazal
    COMPUTATIONAL VISUAL MEDIA, 2024, 10 (6) : 1063 - 1078
  • [48] Disentangling Geometry and Appearance with Regularised Geometry-Aware Generative Adversarial Networks
    Linh Tran
    Jean Kossaifi
    Yannis Panagakis
    Maja Pantic
    International Journal of Computer Vision, 2019, 127 : 824 - 844
  • [49] Few-shot Geometry-Aware Keypoint Localization
    He, Xingzhe
    Bharaj, Gaurav
    Ferman, David
    Rhodin, Helge
    Garrido, Pablo
    2023 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2023, : 21337 - 21348
  • [50] Geometry-Aware Unsupervised Domain Adaptation for Stereo Matching
    Sakuma, Hiroki
    Konishi, Yoshinori
    2021 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION (ICRA 2021), 2021, : 6116 - 6123