Gaussian Mixture Model-Based Registration Network for Point Clouds with Partial Overlap

Li, Xiang; Sun, Jianwen; Own, Chung-Ming; Tao, Wenyuan

doi:10.1007/978-3-031-15934-3_34

Cited by 3 publications

(1 citation statement)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Distributionlevel registration, which compensates for the shortcomings of point-level methods, aligns two point clouds without establishing explicit point correspondences. Unfortunately, to the best of our knowledge, the existing methods are inflexible and cannot handle point clouds with partial overlaps in real scenes [28,31]. Moreover, the success of learningbased methods mainly depends on large amounts of ground truth transformations or correspondences as the supervision signal for model training.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised Deep Probabilistic Approach for Partial Point Cloud Registration

Mei¹,

Tang²,

Huang³

et al. 2023

Preprint

View full text Add to dashboard Cite

Deep point cloud registration methods face challenges to partial overlaps and rely on labeled data. To address these issues, we propose UDPReg, an unsupervised deep probabilistic registration framework for point clouds with partial overlaps. Specifically, we first adopt a network to learn posterior probability distributions of Gaussian mixture models (GMMs) from point clouds. To handle partial point cloud registration, we apply the Sinkhorn algorithm to predict the distribution-level correspondences under the constraint of the mixing weights of GMMs. To enable unsupervised learning, we design three distribution consistencybased losses: self-consistency, cross-consistency, and local contrastive. The self-consistency loss is formulated by encouraging GMMs in Euclidean and feature spaces to share identical posterior distributions. The cross-consistency loss derives from the fact that the points of two partially overlapping point clouds belonging to the same clusters share the cluster centroids. The cross-consistency loss allows the network to flexibly learn a transformation-invariant posterior distribution of two aligned point clouds. The local contrastive loss facilitates the network to extract discriminative local features. Our UDPReg achieves competitive performance on the 3DMatch/3DLoMatch and Model-Net/ModelLoNet benchmarks.

show abstract

Section: Introductionmentioning

confidence: 99%