Filters








48 Hits in 2.3 sec

GVCNN: Group-View Convolutional Neural Networks for 3D Shape Recognition

Yifan Feng, Zizhao Zhang, Xibin Zhao, Rongrong Ji, Yue Gao
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition  
To tackle this issue, in this paper, we propose a group-view convolutional neural network (GVCNN) framework for hierarchical correlation modeling towards discriminative 3D shape description.  ...  However, existing deep features for 3D shape recognition are restricted to a view-to-shape setting, which learns the shape descriptor from the view-level feature directly.  ...  Group-View Convolutional Neural Network In this section, we introduce the proposed GVCNN framework in details.  ... 
doi:10.1109/cvpr.2018.00035 dblp:conf/cvpr/FengZZJG18 fatcat:7oifecut4ffozf3m2ym2vvdnjm

Hypergraph Neural Networks [article]

Yifan Feng, Haoxuan You, Zizhao Zhang, Rongrong Ji, Yue Gao
2019 arXiv   pre-print
We have conducted experiments on citation network classification and visual object recognition tasks and compared HGNN with graph convolutional networks and other traditional methods.  ...  In this paper, we present a hypergraph neural networks (HGNN) framework for data representation learning, which can encode high-order data correlation in a hypergraph structure.  ...  (Su et al. 2015) and Group-View Convolutional Neural Network (GVCNN) (Feng et al. 2018 ).  ... 
arXiv:1809.09401v3 fatcat:uv5kfcjsjbgx7eiqso2bllmupm

Multiple Discrimination and Pairwise CNN for View-based 3D Object Retrieval [article]

Z. Gao, K.X Xue, S.H Wan
2020 arXiv   pre-print
In this work, a novel solution named Multi-view Discrimination and Pairwise CNN (MDPCNN) for 3D object retrieval is proposed to tackle these issues.  ...  However, most existing networks do not take into account the impact of multi-view image selection on network training, and the use of contrastive loss alone only forcing the same-class samples to be as  ...  convolution neural network.  ... 
arXiv:2002.11977v1 fatcat:iq4kii6w2vd7jjbfiwtqmxqcha

Residual Enhanced Multi-Hypergraph Neural Network [article]

Jing Huang, Xiaolin Huang, Jie Yang
2021 arXiv   pre-print
Meanwhile, HyperGraph Neural Network (HGNN) is currently the de-facto method for hypergraph representation learning.  ...  To resolve these issues, we propose the Residual enhanced Multi-Hypergraph Neural Network, which can not only fuse multi-modal information from each hypergraph effectively, but also circumvent the over-smoothing  ...  Multi-View Convolutional Neural Net Network features (MVCNN) [14] and Group-View Convolutional Neural Network (GVCNN) [8] features.  ... 
arXiv:2105.00490v1 fatcat:vndzqkn5xnb5bk33xykf3sar7m

PVCLN: Point-View Complementary Learning Network for 3D Shape Recognition

Shanlin Sun, Yun Li, Minjie Ren, Guo Li, Xing Yao
2020 IEEE Access  
information for 3D shape recognition.  ...  INDEX TERMS 3D shape recognition, point cloud, multiview, complementary learning.  ...  [20] utilized the Hypergraph Neural Network for the 3D shape classification task.  ... 
doi:10.1109/access.2020.3047820 fatcat:6foihghxhbckpevxqhgh7sce5q

Learning Canonical View Representation for 3D Shape Recognition with Arbitrary Views [article]

Xin Wei, Yifei Gong, Fudong Wang, Xing Sun, Jian Sun
2021 arXiv   pre-print
It is a challenging and realistic setting for view-based 3D shape recognition. We propose a canonical view representation to tackle this challenge.  ...  In this way, each 3D shape with arbitrary views is represented by a fixed number of canonical view features, which are further aggregated to generate a rich and robust 3D shape representation for shape  ...  EMV [13] also tries to solve this problem with group convolution over discrete rotation groups.  ... 
arXiv:2108.07084v2 fatcat:q64fvytgpbfkni2yjicjknhlrq

DeepCCFV: Camera Constraint-Free Multi-View Convolutional Neural Network for 3D Object Retrieval

Zhengyue Huang, Zhehui Zhao, Hengguang Zhou, Xibin Zhao, Yue Gao
2019 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
Then, by reducing the over-fitting issue, a camera constraint-free multi-view convolutional neural network named DeepCCFV is constructed.  ...  Most of them require a fixed predefined camera position setting which provides a complete and uniform sampling of views for objects in the training stage.  ...  Since then, more multi-view based neural network methods Kanezaki, Matsushita, and Nishida 2018) have been introduced for 3D shape recognition and retrieval, which assumed a predefined camera setting  ... 
doi:10.1609/aaai.v33i01.33018505 fatcat:oghkzboza5dy7amgaahhnexuta

View-GCN: View-Based Graph Convolutional Network for 3D Shape Analysis

Xin Wei, Ruixuan Yu, Jian Sun
2020 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)  
View-based approach that recognizes 3D shape through its projected 2D images has achieved state-of-the-art results for 3D shape recognition.  ...  In this work, we propose a novel view-based Graph Convolutional Neural Network, dubbed as view-GCN, to recognize 3D shape based on graph representation of multiple views in flexible view configurations  ...  Conclusion We proposed a novel graph convolutional network for 3D shape recognition.  ... 
doi:10.1109/cvpr42600.2020.00192 dblp:conf/cvpr/WeiYS20 fatcat:jdpqrjfgxzg6fjedweh6baqvde

Balanced principal component for 3D shape recognition using convolutional neural networks

Wenjie Luo, Han Zhang, Peng Ni, Xuedong Tian
2020 IET Image Processing  
Currently, PCA (principal component analysis) is widely used in many neural networks and has become a crucial part of the convolutional neural network (CNN) feature extraction.  ...  They addressed 3D shapes using a projection method that is pre-trained on ImageNet and migration learning on ModelNet dataset.  ...  process in convolutional neural network (CNN).  ... 
doi:10.1049/iet-ipr.2019.0844 fatcat:g5v3gch4sbelfb4sftm4la33ia

Gram Regularization for Multi-view 3D Shape Retrieval [article]

Zhaoqun Li
2020 arXiv   pre-print
Most existing 3D shape retrieval methods focus on capturing shape representation with different neural network architectures, while the learning ability of each layer in the network is neglected.  ...  How to obtain the desirable representation of a 3D shape is a key challenge in 3D shape retrieval task.  ...  Introduction 3D shape retrieval is a fundamental research problem in 3D shape analysis, which develops rapidly leveraging the ability of Convolution Neural Networks (CNN).  ... 
arXiv:2011.07733v1 fatcat:bay4tmaoqrhznkeb4n3yy7pfya

PREMA: Part-based REcurrent Multi-view Aggregation Network for 3D Shape Retrieval [article]

Jiongchao Jin, Huanqiang Xu, Pengliang Ji, Zehao Tang, Zhang Xiong
2021 arXiv   pre-print
PREMA accentuates MCPs via correlating features of different views, and aggregates the part-aware features for shape representation.  ...  We propose the Part-based Recurrent Multi-view Aggregation network(PREMA) to eliminate the detrimental effects of the practical view defects, such as insufficient view numbers, occlusions or background  ...  This amounts to a sequential modeling of part-based recognition using recurrent neural networks.  ... 
arXiv:2111.04945v1 fatcat:mlwo3jknkbhnbp2suea35vpvpa

Fusion of a Static and Dynamic Convolutional Neural Network for Multiview 3D Point Cloud Classification

Wenju Wang, Haoran Zhou, Gang Chen, Xiaolin Wang
2022 Remote Sensing  
Therefore, FSDCNet, a neural network model based on the fusion of static and dynamic convolution, is proposed and applied for multiview 3D point cloud classification in this paper.  ...  For the ModelNet40 dataset, the overall accuracy (OA) and average accuracy (AA) of FSDCNet in a single view reached 93.8% and 91.2%, respectively, which were superior to those values for many other methods  ...  [12] presented a group view convolution neural network (GVCNN) for discriminative 3D shape description and hierarchical correlation modeling to better utilize the inherent hierarchical correlation and  ... 
doi:10.3390/rs14091996 fatcat:u6z55mruw5dxjgyej5uylkz36m

ROBUST INDOOR POINT CLOUD CLASSIFICATION BY FUSING LSTM NEURAL NETWORKS WITH SUPERVOXEL CLUSTERING

M. J. Li, L. H. Wang, Z. H. Cai, M. S. Yang, R. J. Wu, M. M. Yao
2022 The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences  
algorithm for indoor point clouds by fusing LSTM neural network and super voxels.  ...  The algorithm first performs super voxel segmentation on the original point cloud and uses it as the basic unit for machine learning classification, and then introduces LSTM (Long Short-Term Memory) neural  ...  GVCNN is a deep neural network for multi-view 3D object recognition, which uses a convolutional neural network with shared view information to extract individual 2D image features for each view, and finally  ... 
doi:10.5194/isprs-archives-xliii-b2-2022-221-2022 doaj:e05139d6ad774667ab42702b97c67c10 fatcat:iok3uqworfesvd3nrrmvoi3dly

Hierarchical Graph Attention based Multi-view Convolutional Neural Network for 3D Object Recognition

Hui Zeng, Tianmeng Zhao, Ruting Cheng, Fuzhou Wang, Jiwei Liu
2021 IEEE Access  
For multi-view convolutional neural network based 3D object recognition, how to fuse the information of multiple views is a key factor affecting the recognition performance.  ...  INDEX TERMS 3D object recognition, multi-view convolutional neural network, graph attention network, feature aggregation. I.  ...  [25] proposed a group-view convolutional neural network named GVCNN, which introduced a grouping module for the neglection of correlation between views. Charles et al.  ... 
doi:10.1109/access.2021.3059853 fatcat:mok35fjqv5exvpaxbekzlj4b6e

MVTN: Multi-View Transformation Network for 3D Shape Recognition [article]

Abdullah Hamdi, Silvio Giancola, Bernard Ghanem
2021 arXiv   pre-print
In particular, we introduce the Multi-View Transformation Network (MVTN) that regresses optimal view-points for 3D shape recognition, building upon advances in differentiable rendering.  ...  As a result, MVTN can be trained end-to-end along with any multi-view network for 3D shape classification.  ...  Gvcnn: Group-view convolutional neural networks for 3d shape recognition. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 264while it is 180 • for MVTN-direct.  ... 
arXiv:2011.13244v3 fatcat:yq2unyuxvfcm3ixjstc5d4vw7a
« Previous Showing results 1 — 15 out of 48 results