Flex-Convolution (Million-Scale Point-Cloud Learning Beyond Grid-Worlds)

Groh, Fabian; Wieschollek, Patrick; Lensch, Hendrik P. A.

doi:10.48550/arxiv.1803.07289

Cited by 11 publications

(14 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Many networks [5], [8], [40] that respect the permutationinvariant property are focused on designing convolution kernels for unordered 3D points. Previous studies [6], [7], [8], [43], [10] reveal effectiveness of considering local geometric details.…”

Section: F Discussionmentioning

confidence: 99%

“…Cires ¸an et al [38] replace subsampling layers with non-overlapping max-pooling layers in the CNNs [39], which achieves surprisingly rapid learning speed and better performance. Various subsampling methods on 3D point clouds [40], [6] have been proposed. Nonetheless, they either do not summarize local geometrical features, or ignore the problem caused by overlapped local neighborhood graphs.…”

Section: B Edge-preserved Poolingmentioning

confidence: 99%

See 1 more Smart Citation

PointAtrousGraph: Deep Hierarchical Encoder-Decoder with Point Atrous Convolution for Unorganized 3D Points

Pan

Chew

Lee

2019

Preprint

View full text Add to dashboard Cite

Motivated by the success of encoding multi-scale contextual information for image analysis, we propose our PointAtrousGraph (PAG) -a deep permutation-invariant hierarchical encoder-decoder for efficiently exploiting multi-scale edge features in point clouds. Our PAG is constructed by several novel modules, such as Point Atrous Convolution (PAC), Edgepreserved Pooling (EP) and Edge-preserved Unpooling (EU). Similar with atrous convolution, our PAC can effectively enlarge receptive fields of filters and thus densely learn multi-scale point features. Following the idea of non-overlapping maxpooling operations, we propose our EP to preserve critical edge features during subsampling. Correspondingly, our EU modules gradually recover spatial information for edge features. In addition, we introduce chained skip subsampling/upsampling modules that directly propagate edge features to the final stage. Particularly, our proposed auxiliary loss functions can further improve our performance. Experimental results show that our PAG outperform previous state-of-the-art methods on various 3D semantic perception applications.

show abstract

Section: F Discussionmentioning

confidence: 99%

Section: B Edge-preserved Poolingmentioning

confidence: 99%

PointAtrousGraph: Deep Hierarchical Encoder-Decoder with Point Atrous Convolution for Unorganized 3D Points

Pan

Chew

Lee

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…However, conventional convolutions cannot be directly applied to point clouds due to the absence of regular grids. Previous networks mostly exploit local point features by two operations: local pooling [24,18,16] and flexible convolution [4,22,13,26]. Self-attention often uses linear layers, such as fully-connected (FC) layers and shared multilayer perceptron (shared MLP) layers, which are appropriate for point clouds.…”

Section: Related Workmentioning

confidence: 99%

Variational Relational Point Completion Network

Pan¹,

Chen²,

Cai³

et al. 2021

Preprint

View full text Add to dashboard Cite

Real-scanned point clouds are often incomplete due to viewpoint, occlusion, and noise. Existing point cloud completion methods tend to generate global shape skeletons and hence lack fine local details. Furthermore, they mostly learn a deterministic partial-to-complete mapping, but overlook structural relations in man-made objects. To tackle these challenges, this paper proposes a variational framework, Variational Relational point Completion network (VRC-Net) with two appealing properties: 1) Probabilistic Modeling. In particular, we propose a dual-path architecture to enable principled probabilistic modeling across partial and complete clouds. One path consumes complete point clouds for reconstruction by learning a point VAE. The other path generates complete shapes for partial point clouds, whose embedded distribution is guided by distribution obtained from the reconstruction path during training. 2) Relational Enhancement. Specifically, we carefully design point selfattention kernel and point selective kernel module to exploit relational point features, which refines local shape de-tails conditioned on the coarse completion. In addition, we contribute a multi-view partial point cloud dataset (MVP dataset) containing over 100,000 high-quality scans, which renders partial 3D shapes from 26 uniformly distributed camera poses for each 3D CAD model. Extensive experiments demonstrate that VRCNet outperforms state-of-theart methods on all standard point cloud completion benchmarks. Notably, VRCNet shows great generalizability and robustness on real-world point cloud scans.

show abstract

“…Point-based Raw point clouds can be taken as input directly without converting them to another regular structured format. It is drawing more and more attention as it enjoys full sparsity [29,26,28,33,34,14,10,39]. Point-Net [26,29] proposes to use shared multi-layer perceptrons and max pooling layers to extract features of point clouds.…”

Section: Related Workmentioning

confidence: 99%

“…Methods OA(%) mAcc(%) Kd-Net [15] 77.4 82.3 SO-Net [19] 81.0 84.9 PCNN by Ext [2] 81.8 85.1 PointNet++ [28] 81.9 85.1 DGCNN [42] 82.3 85.1 SPLATNet [33] 83.7 85.4 Point Convolution KPConv [37] 85.0 86.2 SynSpecCNN [48] 82.0 84.7 Spidercnn [45] 82.4 85.3 SubSparseCNN [8] 83.3 86.0 PointCNN [20] 84.6 86.1 FlexConv [10] 85.0 84.7 PointNet++• P 82.0 85.4 KPConv• P2…”

Section: D Scene Segmentationmentioning

confidence: 99%

Potential Convolution: Embedding Point Clouds into Potential Fields

Chen,

Deng,

et al. 2021

Preprint

View full text Add to dashboard Cite

Recently, various convolutions based on continuous or discrete kernels for point cloud processing have been widely studied, and achieve impressive performance in many applications, such as shape classification, scene segmentation and so on. However, they still suffer from some drawbacks. For continuous kernels, the inaccurate estimation of the kernel weights constitutes a bottleneck for further improving the performance; while for discrete ones, the kernels represented as the points located in the 3D space are lack of rich geometry information. In this work, rather than defining a continuous or discrete kernel, we directly embed convolutional kernels into the learnable potential fields, giving rise to potential convolution. It is convenient for us to define various potential functions for potential convolution which can generalize well to a wide range of tasks. Specifically, we provide two simple yet effective potential functions via point-wise convolution operations. Comprehensive experiments demonstrate the effectiveness of our method, which achieves superior performance on the popular 3D shape classification and scene segmentation benchmarks compared with other state-of-the-art point convolution methods.

show abstract

Flex-Convolution (Million-Scale Point-Cloud Learning Beyond Grid-Worlds)

Cited by 11 publications

References 20 publications

PointAtrousGraph: Deep Hierarchical Encoder-Decoder with Point Atrous Convolution for Unorganized 3D Points

PointAtrousGraph: Deep Hierarchical Encoder-Decoder with Point Atrous Convolution for Unorganized 3D Points

Variational Relational Point Completion Network

Potential Convolution: Embedding Point Clouds into Potential Fields

Contact Info

Product

Resources

About