Multiway Non-Rigid Point Cloud Registration via Learned Functional Map Synchronization

Huang, Jiahui; Birdal, Tolga; Gojčič, Žan; Guibas, Leonidas J.; Hu, Shi‐Min

doi:10.1109/tpami.2022.3164653

Cited by 20 publications

(7 citation statements)

References 74 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, deep learning based methods also attract attentions [16][17][18][19]41]. For example, Shimada et al [19] predicted the displacement fields on a voxel grid by the Displacements on Voxels Networks (DispVoxNets); Feng et al [17] represented the non-rigid transformation with a combination of K rigid transformations and, accordingly, proposed a GRU-based framework.…”

Section: Non-rigid Registrationmentioning

confidence: 99%

See 1 more Smart Citation

Automatic Extraction of the Sparse Prior Correspondences for Non-Rigid Point Cloud Registration

Zhu¹,

Tian²,

Ye³

et al. 2023

Computer Modeling in Engineering &Amp; Sciences

View full text Add to dashboard Cite

Non-rigid registration of point clouds is still far from stable, especially for the largely deformed one. Sparse initial correspondences are often adopted to facilitate the process. However, there are few studies on how to build them automatically. Therefore, in this paper, we propose a robust method to compute such priors automatically, where a global and local combined strategy is adopted. These priors in different degrees of deformation are obtained by the locally geometrical-consistent point matches from the globally structural-consistent region correspondences. To further utilize the matches, this paper also proposes a novel registration method based on the Coherent Point Drift framework. This method takes both the spatial proximity and local structural consistency of the priors as supervision of the registration process and thus obtains a robust alignment for clouds with significantly different deformations. Qualitative and quantitative experiments demonstrate the advantages of the proposed method.

show abstract

Section: Non-rigid Registrationmentioning

confidence: 99%

“…In addition, in comparison with the recent developments in deep neural networks based ideas [16][17][18][19], our method belongs to the traditionally artificial feature based methods which is easier and more stable to apply without a big training set and tedious parametric tuning.…”

Section: Introductionmentioning

confidence: 99%

Automatic Extraction of the Sparse Prior Correspondences for Non-Rigid Point Cloud Registration

Zhu¹,

Tian²,

Ye³

et al. 2023

Computer Modeling in Engineering &Amp; Sciences

View full text Add to dashboard Cite

show abstract

“…Besides N-ICP, RPTS and SVR-ℓ 0 , we also compare with four recent deep learning methods, Lepard+NICP [65], NDP [66], RMA-Net [44], and SyNoRiM [67] using their open-source implementations 3,4,5 . For each frame, we first downsample the depth map and obtain a point cloud according to the intrinsic camera parameters, then remove the background using the ground-truth segmentation masks provided by the dataset.…”

Section: Datamentioning

confidence: 99%

Fast and Robust Non-Rigid Registration Using Accelerated Majorization-Minimization

Yao

Deng

et al. 2023

IEEE Trans. Pattern Anal. Mach. Intell.

View full text Add to dashboard Cite

Non-rigid 3D registration, which deforms a source 3D shape in a non-rigid way to align with a target 3D shape, is a classical problem in computer vision. Such problems can be challenging because of imperfect data (noise, outliers and partial overlap) and high degrees of freedom. Existing methods typically adopt the ℓp type robust norm to measure the alignment error and regularize the smoothness of deformation, and use a proximal algorithm to solve the resulting non-smooth optimization problem. However, the slow convergence of such algorithms limits their wide applications. In this paper, we propose a formulation for robust non-rigid registration based on a globally smooth robust norm for alignment and regularization, which can effectively handle outliers and partial overlaps. The problem is solved using the majorization-minimization algorithm, which reduces each iteration to a convex quadratic problem with a closed-form solution. We further apply Anderson acceleration to speed up the convergence of the solver, enabling the solver to run efficiently on devices with limited compute capability. Extensive experiments demonstrate the effectiveness of our method for non-rigid alignment between two shapes with outliers and partial overlaps, with quantitative evaluation showing that it outperforms state-of-the-art methods in terms of registration accuracy and computational speed. The source code is available at https://github.com/yaoyx689/AMM NRR.

show abstract

“…Caspr [47] implicitly accumulates the shapes by mapping a sequence of partial observations to a continuous latent space. [24] and [23] respectively propose multi-way registration methods that accumulate multi-body and non-rigid dynamic point clouds, but do not scale well to large scenes.…”

Section: Related Workmentioning

confidence: 99%

“…Starting from these observations, we propose a novel point cloud accumulation scheme tailored to the autonomous driving setting. To that end, we aim to accumulate point clouds over time while abstracting the scene into a collection of rigidly moving agents [3,19,52] and reasoning about each agent's motion on the basis of a longer sequence of frames [24,23]. Along with the accumulated point cloud (Fig.…”

mentioning

confidence: 99%

Dynamic 3D Scene Analysis by Point Cloud Accumulation

Huang¹,

Gojčič²,

Huang³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Multiway Non-Rigid Point Cloud Registration via Learned Functional Map Synchronization

Cited by 20 publications

References 74 publications

Automatic Extraction of the Sparse Prior Correspondences for Non-Rigid Point Cloud Registration

Automatic Extraction of the Sparse Prior Correspondences for Non-Rigid Point Cloud Registration

Fast and Robust Non-Rigid Registration Using Accelerated Majorization-Minimization

Dynamic 3D Scene Analysis by Point Cloud Accumulation

Contact Info

Product

Resources

About