Filters








7,628 Hits in 8.7 sec

From Points to Parts: 3D Object Detection from Point Cloud with Part-aware and Part-aggregation Network [article]

Shaoshuai Shi, Zhe Wang, Jianping Shi, Xiaogang Wang, Hongsheng Li
2020 arXiv   pre-print
In this paper, we extend our preliminary work PointRCNN to a novel and strong point-cloud-based 3D object detection framework, the part-aware and aggregation neural network (Part-A^2 net).  ...  3D object detection from LiDAR point cloud is a challenging problem in 3D scene understanding and has many practical applications.  ...  With such rich supervisions, we propose a novel part-aware and aggregation 3D object detector, Part-A 2 net, for 3D object detection from point cloud.  ... 
arXiv:1907.03670v3 fatcat:ungdwipuwzhprpd2ymgcir3cdy

Non-Local Part-Aware Point Cloud Denoising [article]

Chao Huang, Ruihui Li, Xianzhi Li, Chi-Wing Fu
2020 arXiv   pre-print
This paper presents a novel non-local part-aware deep neural network to denoise point clouds by exploring the inherent non-local self-similarity in 3D objects and scenes.  ...  features over the entire point cloud.  ...  denoised point cloud: L part = M i=1 CD(q i ,q i ), (3) where M is the total number of parts in the object (which varies from objects to objects).  ... 
arXiv:2003.06631v1 fatcat:o7gjmwbdxndj5btadq37ite6uu

PartGlot: Learning Shape Part Segmentation from Language Reference Games [article]

Juil Koo, Ian Huang, Panos Achlioptas, Leonidas Guibas, Minhyuk Sung
2022 arXiv   pre-print
, essential to their recognition and use.  ...  We introduce PartGlot, a neural framework and associated architectures for learning semantic part segmentation of 3D shape geometry, based solely on part referential language.  ...  We test our network (with the PN-Aware setup) while varying the N from 16 to 256 (and 2048, which is the extreme case).  ... 
arXiv:2112.06390v2 fatcat:wvpdbuengjbhphnqfyvzotfvcu

Learning 3D Part Assembly from a Single Image [article]

Yichen Li and Kaichun Mo and Lin Shao and Minhyuk Sung and Leonidas Guibas
2020 arXiv   pre-print
We study this problem in the setting of furniture assembly from a given complete set of parts and a single image depicting the entire assembled object.  ...  We address these issues by proposing a two-module pipeline that leverages strong 2D-3D correspondences and assembly-oriented graph message-passing to infer part relationships.  ...  First, the network needs to distinguish between the geometric subtlety of the input part point clouds to establish a valid 2D-3D correspondence.  ... 
arXiv:2003.09754v2 fatcat:gnjqkoimifel7c4kz5efpclmfq

Generative 3D Part Assembly via Dynamic Graph Learning [article]

Jialei Huang, Guanqi Zhan, Qingnan Fan, Kaichun Mo, Lin Shao, Baoquan Chen, Leonidas Guibas, Hao Dong
2020 arXiv   pre-print
It explicitly conducts sequential part assembly refinements in a coarse-to-fine manner, exploits a pair of part relation reasoning module and part aggregation module for dynamically adjusting both part  ...  Essentially, the task of generative 3D part assembly is to predict a 6-DoF part pose, including a rigid rotation and translation, for each input part that assembles a single 3D shape as the final output  ...  Structure-aware generative networks.  ... 
arXiv:2006.07793v3 fatcat:nitzidsvnnaihmhanlni7vzwqq

Act the Part: Learning Interaction Strategies for Articulated Object Part Discovery [article]

Samir Yitzhak Gadre, Kiana Ehsani, Shuran Song
2021 arXiv   pre-print
To this end, we introduce Act the Part (AtP) to learn how to interact with articulated objects to discover and segment their pieces.  ...  Motivated by this observation, we identify an important embodied task where an agent must play with objects to recover their parts.  ...  This work was supported in part by the Amazon Research Award and NSF CMMI-2037101.  ... 
arXiv:2105.01047v1 fatcat:kh343adl25f5zehv2t2h7nvfjq

ZoomNet: Part-Aware Adaptive Zooming Neural Network for 3D Object Detection [article]

Zhenbo Xu, Wei Zhang, Xiaoqing Ye, Xiao Tan, Wei Yang, Shilei Wen, Errui Ding, Ajin Meng, Liusheng Huang
2020 arXiv   pre-print
In this way, we are able to estimate higher-quality disparity maps from the resized box images then construct dense point clouds for both nearby and distant objects.  ...  Moreover, we introduce to learn part locations as complementary features to improve the resistance against occlusion and put forward the 3D fitting score to better estimate the 3D detection quality.  ...  Acknowledgment This work was supported by the National Natural Science Foundation of China (No. 61602400) and the Anhui Initiative in Quantum Information Technologies (No. AHY150300).  ... 
arXiv:2003.00529v1 fatcat:kfcz5eoqvnaflhhlfqt6mkeb7a

ZoomNet: Part-Aware Adaptive Zooming Neural Network for 3D Object Detection

Zhenbo Xu, Wei Zhang, Xiaoqing Ye, Xiao Tan, Wei Yang, Shilei Wen, Errui Ding, Ajin Meng, Liusheng Huang
2020 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
In this way, we are able to estimate higher-quality disparity maps from the resized box images then construct dense point clouds for both nearby and distant objects.  ...  Moreover, we introduce to learn part locations as complementary features to improve the resistance against occlusion and put forward the 3D fitting score to better estimate the 3D detection quality.  ...  Acknowledgment This work was supported by the National Natural Science Foundation of China (No. 61602400) and the Anhui Initiative in Quantum Information Technologies (No. AHY150300).  ... 
doi:10.1609/aaai.v34i07.6945 fatcat:wngiiy3npvdxpnxvz5vzjassnm

OPD: Single-view 3D Openable Part Detection [article]

Hanxiao Jiang, Yongsen Mao, Manolis Savva, Angel X. Chang
2022 arXiv   pre-print
To tackle this task, we create two datasets of 3D objects: OPDSynth based on existing synthetic objects, and OPDReal based on RGBD reconstructions of real objects.  ...  The input is a single image of an object, and as output we detect what parts of the object can open, and the motion parameters describing the articulation of each openable part.  ...  Acknowledgement: This work was funded in part by a Canada CIFAR AI Chair, a Canada Research Chair and NSERC Discovery Grant, and enabled in part by support from WestGrid and Compute Canada.  ... 
arXiv:2203.16421v1 fatcat:v2esge3hxrbrbewejqt5gnrpui

Guest Editorial Spectrum Sharing and Aggregation for Future Wireless Networks, Part III

Theodoros A. Tsiftsis, Guoru Ding, Yulong Zou, George K. Karagiannidis, Zhu Han, Lajos Hanzo
2017 IEEE Journal on Selected Areas in Communications  
The eighth paper, which is entitled as "Efficient 3D resource management for spectrum aggregation in cellular networks," introduces the power domain for spectrum aggregation from the perspective of the  ...  This work also investigates an underlay spectrum sharing regime between the 3D DSC networks and the traditional cellular networks modeled by 2D Poisson point processes.  ... 
doi:10.1109/jsac.2016.2633098 fatcat:dhdkkqr5knhuzhx5an7vhspsda

Learning Local Shape Descriptors from Part Correspondences With Multi-view Convolutional Networks [article]

Haibin Huang, Evangelos Kalogerakis, Siddhartha Chaudhuri, Duygu Ceylan, Vladimir G. Kim, Ersin Yumer
2017 arXiv   pre-print
We present a new local descriptor for 3D shapes, directly applicable to a wide range of shape analysis problems such as point correspondences, semantic segmentation, affordance prediction, and shape-to-scan  ...  The descriptor is produced by a convolutional network that is trained to embed geometrically and semantically similar points close to one another in descriptor space.  ...  ACKNOWLEDGEMENTS Kalogerakis acknowledges support from NSF (CHS-1422441, CHS-1617333), NVidia and Adobe. Chaudhuri acknowledges support from Adobe and Qualcomm.  ... 
arXiv:1706.04496v2 fatcat:26jeic2jb5gx7l4aonrdqak7di

Spatial-importance based computation scheme for real-time object detection from 3D sensor data

Ryo Otsu, Ryoichi Shinkuma, Takehiro Sato, Eiji Oki, Daiki Hasegawa, Toshikazu Furuya
2022 IEEE Access  
Three-dimensional (3D) sensor networks using multiple light-detection-and-ranging (LIDAR) sensors are good for smart monitoring of spots, such as intersections, with high potential risk of roadtraffic  ...  It processes point-cloud data from each region according to the spatial importance of that region.  ...  The part-aggregation stage conducts RoI-aware point cloud pooling to aggregate the object part information for each 3D proposal.  ... 
doi:10.1109/access.2022.3140332 fatcat:4zvlhckkb5g6bghrddsxsm3a3i

Task-Aware Sampling Layer for Point-Wise Analysis [article]

Yiqun Lin, Lichang Chen, Haibin Huang, Chongyang Ma, Xiaoguang Han, Shuguang Cui
2022 arXiv   pre-print
We further demonstrate our methods in various point-wise analysis tasks, including semantic part segmentation, point cloud completion, and keypoint detection.  ...  Sampling, grouping, and aggregation are three important components in the multi-scale analysis of point clouds.  ...  The displacement network is supervised by task-aware sampling with CD (or EMD) loss, and also jointly supervised with task loss since it is one part of the whole task network.  ... 
arXiv:2107.04291v3 fatcat:psv3ao3v2bgltfyuxcwyypiuoe

Box-Aware Feature Enhancement for Single Object Tracking on Point Clouds [article]

Chaoda Zheng, Xu Yan, Jiantao Gao, Weibing Zhao, Wei Zhang, Zhen Li, Shuguang Cui
2021 arXiv   pre-print
In particular, we first propose the BoxCloud, an informative and robust representation, to depict an object using the point-to-box relation.  ...  Current 3D single object tracking approaches track the target based on a feature comparison between the target template and the search area.  ...  In 3D detection, [30] predicts intra-object part locations of all 3D points within a proposal, using the freeof-charge part supervisions derived from 3D ground-truth BBoxes.  ... 
arXiv:2108.04728v2 fatcat:ixvy3dbwnrfkxppirwgiubovgq

Deep Learning for 3D Point Clouds: A Survey [article]

Yulan Guo, Hanyun Wang, Qingyong Hu, Hao Liu, Li Liu, Mohammed Bennamoun
2020 arXiv   pre-print
It covers three major tasks, including 3D shape classification, 3D object detection and tracking, and 3D point cloud segmentation.  ...  However, deep learning on point clouds is still in its infancy due to the unique challenges faced by the processing of point clouds with deep neural networks.  ...  [125] proposed a graph neural network Point-GNN to detect 3D objects from lidar point clouds.  ... 
arXiv:1912.12033v2 fatcat:qiiyvvuulfccxaiihf2mu23k34
« Previous Showing results 1 — 15 out of 7,628 results