PREDATOR: Registration of 3D Point Clouds with Low Overlap

Huang, Shengyu; Gojčič, Žan; Usvyatsov, Mikhail; Wieser, Andreas; Schindler, Konrad

doi:10.1109/cvpr46437.2021.00425

Cited by 331 publications

(333 citation statements)

References 26 publications

Supporting

Mentioning

285

Contrasting

Order By: Relevance

“…As shown in Table 10, our method consistently outperforms RANSAC when combined with different descriptors. Moreover, our method can further boost the performance of Predator [32], a recently proposed learning-based descriptors especially designed for low-overlapping registration, showing the effectiveness and robustness of our method under high outlier ratios. PointDSC increases the registration recall by 16.3% and 7.3% under 5000 points setting for FCGF and Predator, respectively.…”

Section: Additional Experimentsmentioning

confidence: 79%

“…Traditional point cloud registration algorithms (e.g., [8,1,50,33,46,46]) have been comprehensively reviewed in [56]. Recently, learningbased algorithms have been proposed to replace the individual components in the classical registration pipeline, including keypoint detection [4,40,34] and feature description [21,22,23,55,4,18,28,32,2]. Besides, end-toend registration networks [3,67,68,76] have been proposed.…”

Section: Related Workmentioning

confidence: 99%

“…The computation of registration recall is slightly different with ours, we refer readers to[32] for more details.…”

mentioning

confidence: 99%

See 2 more Smart Citations

PointDSC: Robust Point Cloud Registration using Deep Spatial Consistency

Bai¹,

Luo²,

Zhou³

et al. 2021

Preprint

View full text Add to dashboard Cite

Removing outlier correspondences is one of the critical steps for successful feature-based point cloud registration. Despite the increasing popularity of introducing deep learning techniques in this field, spatial consistency, which is essentially established by a Euclidean transformation between point clouds, has received almost no individual attention in existing learning frameworks. In this paper, we present PointDSC, a novel deep neural network that explicitly incorporates spatial consistency for pruning outlier correspondences. First, we propose a nonlocal feature aggregation module, weighted by both feature and spatial coherence, for feature embedding of the input correspondences. Second, we formulate a differentiable spectral matching module, supervised by pairwise spatial compatibility, to estimate the inlier confidence of each correspondence from the embedded features. With modest computation cost, our method outperforms the state-of-the-art handcrafted and learning-based outlier rejection approaches on several real-world datasets by a significant margin. We also show its wide applicability by combining PointDSC with different 3D local descriptors.[code release]

show abstract

Section: Additional Experimentsmentioning

confidence: 79%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

PointDSC: Robust Point Cloud Registration using Deep Spatial Consistency

Bai¹,

Luo²,

Zhou³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…The recent advances have been dominated by learningbased, correspondence-based methods [4,8,10,14,15,39]. A neural network is trained to extract point correspondences between two input point clouds, based on which an alignment transformation is calculated with a robust estimator, e.g., RANSAC.…”

Section: Introductionmentioning

confidence: 99%

“…A neural network is trained to extract point correspondences between two input point clouds, based on which an alignment transformation is calculated with a robust estimator, e.g., RANSAC. Most correspondence-based methods rely on keypoint detection [1,4,8,15]. However, it is challenging to detect repeatable keypoints across two point clouds, especially when they have small overlapping area.…”

Section: Introductionmentioning

confidence: 99%

Geometric Transformer for Fast and Robust Point Cloud Registration

Zheng¹,

Yu²,

Wang³

et al. 2022

Preprint

View full text Add to dashboard Cite

We study the problem of extracting accurate correspondences for point cloud registration. Recent keypoint-free methods bypass the detection of repeatable keypoints which is difficult in low-overlap scenarios, showing great potential in registration. They seek correspondences over downsampled superpoints, which are then propagated to dense points. Superpoints are matched based on whether their neighboring patches overlap. Such sparse and loose matching requires contextual features capturing the geometric structure of the point clouds. We propose Geometric Transformer to learn geometric feature for robust superpoint matching. It encodes pair-wise distances and triplet-wise angles, making it robust in low-overlap cases and invariant to rigid transformation. The simplistic design attains surprisingly high matching accuracy such that no RANSAC is required in the estimation of alignment transformation, leading to 100 times acceleration. Our method improves the inlier ratio by 17%∼30% and the registration recall by over 7% on the challenging 3DLoMatch benchmark. The code and models will be released at https://github. com/qinzheng93/GeoTransformer.

show abstract

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

Sun

Own

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

PREDATOR: Registration of 3D Point Clouds with Low Overlap

Cited by 331 publications

References 26 publications

PointDSC: Robust Point Cloud Registration using Deep Spatial Consistency

PointDSC: Robust Point Cloud Registration using Deep Spatial Consistency

Geometric Transformer for Fast and Robust Point Cloud Registration

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

Contact Info

Product

Resources

About