Point clouds learning with attention-based graph convolution networks

Xie, Zhuyang; Chen, Junzhou; Peng, Bo

doi:10.1016/j.neucom.2020.03.086

Cited by 56 publications

(36 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 1 shows the classification results of the proposed DNet compared with the other sixteen advanced networks. As shown in the “input” column of Table 1 , the methods, including the Spec-GCN [ 15 ], Pointconv [ 6 ], AGCN [ 41 ], PointNet++ [ 21 ], SpiderCNN [ 7 ] and SO-Net [ 28 ], require coordinates of point cloud as well as normal information as the input of their networks, while the other eleven comparison networks and the proposed DNet only need the coordinates of point cloud. Moreover, the networks listed in the last three (PointNet++ [ 21 ], SpiderCNN [ 7 ] and SO-Net [ 28 ]) for comparison use 5k points, rather than 1k points as other networks do.…”

Section: Resultsmentioning

confidence: 99%

“…Attention mechanism was used for weighting aggregation of point features in local regions [ 17 , 39 , 40 , 41 ], and it is also important for neighborhood learning. Chen et al [ 17 ] used graph attention mechanism to learn local geometric representations of point clouds.…”

Section: Motivationmentioning

confidence: 99%

“…Feng et al [ 40 ] proposed a Local Attention-Edge Convolution (LAE-Conv) to construct a local graph based on the neighborhood points searched in multi-directions. Xie et al [ 41 ] used the local graph structure and the global graph structure to enhance the feature learning of point clouds. However, the traditional attention mechanism mainly focuses on using different features to obtain the weights of the neighborhood points, even for the pathological neighborhood as shown in Figure 1 , such a kind of attention network also counts these pathological neighborhood points.…”

Section: Motivationmentioning

confidence: 99%

See 2 more Smart Citations

DNet: Dynamic Neighborhood Feature Learning in Point Cloud

Tian

Jiang

2021

Sensors

View full text Add to dashboard Cite

Neighborhood selection is very important for local region feature learning in point cloud learning networks. Different neighborhood selection schemes may lead to quite different results for point cloud processing tasks. The existing point cloud learning networks mainly adopt the approach of customizing the neighborhood, without considering whether the selected neighborhood is reasonable or not. To solve this problem, this paper proposes a new point cloud learning network, denoted as Dynamic neighborhood Network (DNet), to dynamically select the neighborhood and learn the features of each point. The proposed DNet has a multi-head structure which has two important modules: the Feature Enhancement Layer (FELayer) and the masking mechanism. The FELayer enhances the manifold features of the point cloud, while the masking mechanism is used to remove the neighborhood points with low contribution. The DNet can learn the manifold features and spatial geometric features of point cloud, and obtain the relationship between each point and its effective neighborhood points through the masking mechanism, so that the dynamic neighborhood features of each point can be obtained. Experimental results on three public datasets demonstrate that compared with the state-of-the-art learning networks, the proposed DNet shows better superiority and competitiveness in point cloud processing task.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Motivationmentioning

confidence: 99%

Section: Motivationmentioning

confidence: 99%

See 1 more Smart Citation

DNet: Dynamic Neighborhood Feature Learning in Point Cloud

Tian

Jiang

2021

Sensors

View full text Add to dashboard Cite

show abstract

“…This model projects the features of disordered points onto the ordered sequence of feature vectors, which are then fed into the recurrent neural network for learning. Besides, Graph Convolutional Neural Networks [30][31][32] and Multi-scale Graph Convolutional Neural Networks [15,33] also are used to deal with irregular data structure. Mao et al [16] and Thomas et al [17] present new convolution methods to extract local features.…”

Section: End-to-end Learning Networkmentioning

confidence: 99%

KVGCN: A KNN Searching and VLAD Combined Graph Convolutional Network for Point Cloud Segmentation

Luo

Huo

et al. 2021

Remote Sensing

View full text Add to dashboard Cite

Semantic segmentation of the sensed point cloud data plays a significant role in scene understanding and reconstruction, robot navigation, etc. This work presents a Graph Convolutional Network integrating K-Nearest Neighbor searching (KNN) and Vector of Locally Aggregated Descriptors (VLAD). KNN searching is utilized to construct the topological graph of each point and its neighbors. Then, we perform convolution on the edges of constructed graph to extract representative local features by multiple Multilayer Perceptions (MLPs). Afterwards, a trainable VLAD layer, NetVLAD, is embedded in the feature encoder to aggregate the local and global contextual features. The designed feature encoder is repeated for multiple times, and the extracted features are concatenated in a jump-connection style to strengthen the distinctiveness of features and thereby improve the segmentation. Experimental results on two datasets show that the proposed work settles the shortcoming of insufficient local feature extraction and promotes the accuracy (mIoU 60.9% and oAcc 87.4% for S3DIS) of semantic segmentation comparing to existing models.

show abstract

“…By constructing a graph on a point cloud, the graph convolution operation can extract the local feature of a point and the points around it in the graph. Both [6] and [12] have constructed a graph before applying graph convolution, however, they only learn the information from the points nearby. It needs the networks to downsample the point clouds for learning global semantic information.…”

Section: Introductionmentioning

confidence: 99%

Multi-Branch Global Graph Convolution Network for Point Clouds

Fan

Liu

et al. 2021

IEEE Access

View full text Add to dashboard Cite

For finding the defect inside of rails, we usually use the ultrasound to inspect the rails and obtain point clouds data of the returned ultrasonic signals after preprocessing. The quantity of points in each point cloud is not fixed, and the point clouds are disordered and unstructured. The points have heterogeneous attributes which include not only continuous coordinates but discrete frequencies of the ultrasound. For classifying the special point clouds, we propose a network architecture, named Multi-Branch Global Graph Convolution Network. For better utilizing the features of heterogeneous attributes of each point, we introduce the point channel-wise attention mechanism to weight the attributes. The backbone of the network is a multi-branch learn network based on the global graph convolution. By constructing the global graph and applying the graph convolution, the network possesses the ability to learn the local relative position information and global semantic information. Experimental results show the effectiveness of each component of the network and the state-of-the-art performance of the proposed network achieves on the classification task of defect recognition.

show abstract

Point clouds learning with attention-based graph convolution networks

Cited by 56 publications

References 31 publications

DNet: Dynamic Neighborhood Feature Learning in Point Cloud

DNet: Dynamic Neighborhood Feature Learning in Point Cloud

KVGCN: A KNN Searching and VLAD Combined Graph Convolutional Network for Point Cloud Segmentation

Multi-Branch Global Graph Convolution Network for Point Clouds

Contact Info

Product

Resources

About