Robust Semi-Supervised Point Cloud Registration via Latent GMM-Based Correspondence

Zhang, Zhengyan1; Lyu, Erli2; Min, Zhe3; Zhang, Ang4; Yu, Yue5; Meng, Max Q. -H.6

2023-09-01

10.3390/rs15184493

REMOTE SENSING   影响因子和分区

2072-4292

Due to the fact that point clouds are always corrupted by significant noise and large transformations, aligning two point clouds by deep neural networks is still challenging. This paper presents a semi-supervised point cloud registration (PCR) method for accurately estimating point correspondences and handling large transformations using limited prior datasets. Firstly, a modified autoencoder is introduced as the feature extraction module to extract the distinctive and robust features for the downstream registration task. Unlike optimization-based pairwise PCR strategies, the proposed method treats two point clouds as two implementations of a Gaussian mixture model (GMM), which we call latent GMM. Based on the above assumption, two point clouds can be regarded as two probability distributions. Hence, the PCR of two point clouds can be approached by minimizing the KL divergence between these two probability distributions. Then, the correspondence between the point clouds and the latent GMM components is estimated using the augmented regression network. Finally, the parameters of GMM can be updated by the correspondence and the transformation matrix can be computed by employing the weighted singular value decomposition (SVD) method. Extensive experiments conducted on both synthetic and real-world data validate the superior performance of the proposed method compared to state-of-the-art registration methods. These experiments also highlight the method's superiority in terms of accuracy, robustness, and generalization.

point cloud registration deep learning methods Gaussian mixture model semi-supervised learning

SCI ; EI

英语

其他

Environmental Sciences & Ecology ; Geology ; Remote Sensing ; Imaging Science & Photographic Technology

Environmental Sciences ; Geosciences, Multidisciplinary ; Remote Sensing ; Imaging Science & Photographic Technology

WOS:001074395700001

MDPI

20234014835081

Gaussian distribution ; Large dataset ; Learning systems ; Linear transformations ; Singular value decomposition

Ergonomics and Human Factors Engineering:461.4 ; Data Processing and Image Processing:723.2 ; Mathematics:921 ; Mathematical Transformations:921.3 ; Probability Theory:922.1 ; Mathematical Statistics:922.2

Web of Science

1.Harbin Inst Technol, Sch Mech Engn & Automat, Shenzhen 02138, Peoples R China
2.Macao Polytech Univ, Fac Appl Sci, Macau 999078, Peoples R China
3.UCL, Dept Med Phys & Biomed Engn, London WC1E 6BT, England
4.Chinese Univ Hong Kong, Dept Elect Engn, Hong Kong 999077, Peoples R China
5.Wuhan Univ, State Key Lab Informat Engn Surveying Mapping & Re, Wuhan 430079, Peoples R China
6.Southern Univ Sci & Technol, Dept Elect & Elect Engn, Shenzhen 518055, Peoples R China

Zhang, Zhengyan,Lyu, Erli,Min, Zhe,et al. Robust Semi-Supervised Point Cloud Registration via Latent GMM-Based Correspondence[J]. REMOTE SENSING,2023,15(18).

Zhang, Zhengyan,Lyu, Erli,Min, Zhe,Zhang, Ang,Yu, Yue,&Meng, Max Q. -H..(2023).Robust Semi-Supervised Point Cloud Registration via Latent GMM-Based Correspondence.REMOTE SENSING,15(18).

Zhang, Zhengyan,et al."Robust Semi-Supervised Point Cloud Registration via Latent GMM-Based Correspondence".REMOTE SENSING 15.18(2023).