一种面向低采样率的点云数据处理网络

doi:10.19678/j.issn.1000-3428.0063204

摘要/Abstract

摘要： 在直接处理点云的三维神经网络中，采样阶段实现了对原始点云中关键点的筛选，对于整个网络的性能及网络的抗噪能力具有重要作用。目前主流的最远点采样（FPS）方法在处理大规模3D点云数据时计算量大且耗时，并且低采样率时经过FPS采样后模型性能下降明显。针对这两个问题，提出一种面向低采样率的点云数据处理网络AS-Net。设计一个新的采样模块代替原backbone中的FPS，其由两个Layer组成，每个Layer基于长短期记忆网络获取原始点云与采样点云之间的联系权重，从而高效提取关键信息，去除冗余信息。在此基础上，利用注意力机制选择特征值较高的原始点云作为采样点，采样点作为后序任务的关键点输入到网络，进一步提高网络模型性能。基于ModelNet40数据集的实验结果表明，在低采样率条件下，AS-Net仍可达到81.6%的分类准确率，与使用FPS作为采样方法的网络模型相比提高52.7%。此外，其对噪声干扰具有很强的鲁棒性，对于大场景的分割时间效率优于同类采样方法。

关键词: 点云, 采样, 最远点采样, 长短期记忆网络, 注意力机制

Abstract: Point-cloud data sampling, grouping and fusion have been applied in several state-of-the-art networks to improve their segmentation and classification results.Among these, the sampling stage maintains the key points of the original point cloud, which plays an important role in the performance and anti-noise ability of the network.However, two defects limit the application of models based on Farthest Point Sampling(FPS) to large-scale 3D point-cloud data:(1) FPS requires considerable computation and time;(2) at low sampling rates, the performance of the model under FPS significantly decreases.To solve those two problems, a point cloud data processing network oriented to low sampling rate, AS-Net, is proposed, which uses a new sampling module to replace FPS in other backbones.The sampling module is composed of two layers:(1) a Long Short-Term Memory(LSTM) network structure;(2) an attention mechanism.LSTM is used as the feature extraction method to obtain the relational weight relative to the subsequent tasks, efficiently extract key information, and remove redundant information.The attention mechanism is used to select the original point cloud with a high eigenvalue point as the sampling point;sampling points are input to the network as key points of the subsequent tasks to improve the performance of the network model.Experiments show that the proposed method still achieves a classification accuracy of 81.6% on the ModelNet40 dataset at a low sampling rate, which is 52.7% higher than that of the same network using FPS.In addition, the proposed method has a strong robustness to noise interference, and its segmentation time efficiency for large scenes is better than that of similar algorithms.

Key words: point cloud, sampling, Farthest Point Sampling(FPS), Long Short-Term Memory(LSTM) network, attention mechanism

中图分类号:

TP391.41

张毅, 林云汉, 刘双元. 一种面向低采样率的点云数据处理网络[J]. 计算机工程, 2022, 48(11): 240-246.

ZHANG Yi, LIN Yunhan, LIU Shuangyuan. A Point Cloud Data Processing Network Oriented to Low Sampling Rate[J]. Computer Engineering, 2022, 48(11): 240-246.

https://www.ecice06.com/CN/Y2022/V48/I11/240

图/表 11

20230221180645

20230221180648

20230221180652

20230221180655

20230221180658

20230221180701

20230221180705

20230221180708

20230221180711

20230221180715

20230221180719

参考文献

[1] CHEN X, MA H, WAN J, et al.Multi-view 3D object detection network for autonomous[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C, USA:IEEE Press, 2017:652-660.
[2] GEIGER A, LENZ P, URTASUN R, et al.Are we ready for autonomous driving? The KITTI vision benchmark suite[C]//Proceedings of 2012 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C, USA:IEEE Press, 2012:3354-3361.
[3] MARCHAND E, UCHIYAMA H, SPINDLER F.Pose estimation for augmented reality:a hands-on survey[J].IEEE Transactions on Visualization and Computer Graphics, 2016, 22(12):2633-2651.
[4] 钟诚, 周浩杰, 韦海亮.一种基于注意力机制的三维点云物体识别方法[J].计算机技术与发展, 2020, 30(4):41-45. ZHONG C, ZHOU H J, WEI H L.A 3D point cloud object recognition method based on attention mechanism[J].Computer Technology and Development, 2020, 30(4):41-45.(in Chinese)
[5] XU M, DING R, ZHAO H, et al.PAConv:position adaptive convolution with dynamic kernel assembling on point clouds[C]//Proceedings of 2021 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C, USA:IEEE Press, 2021:3173-3182.
[6] COLLET A, MARTINEZ M, SRINIVASA S S.The MOPED framework:object recognition and pose estimation for manipulation[J].The International Journal of Robotics Research, 2011, 30(10):1284-1306.
[7] TREMBLAY J, TO T, SUNDARALINGAM B, et al.Deep object pose estimation for semantic robotic grasping of household objects[EB/OL].[2021-10-11].https://arxiv.org/abs/1809.10790.
[8] QI C.R, YI L, SU H, GUIBAS L J.PointNet++:deep hierarchical feature learning on point sets in a metric space[C]//Proceedings of 2017 Conference on Neural Information Processing Systems.Cambridge, USA:MIT Press, 2017:5099-5108.
[9] LIU Y, FAN B, XIANG S, et al.Relation-shape convolutional neural network for point cloud analysis[C]//Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C, USA:IEEE Press, 2019:8895-8904.
[10] HU Q, YANG B, XIE L, et al.RandLA-Net:efficient semantic segmentation of large-scale point clouds[C]//Proceedings of 2019 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C, USA:IEEE Press, 2019:11108-11117.
[11] XU D, ANGUELOV D, JAIN A, et al.PointFusion:deep sensor fusion for 3D bounding box estimation[C]//Proceedings of 2018 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C, USA:IEEE Press, 2018:244-253.
[12] CHEN X, MA H, WAN J, et al.Multi-view 3D object detection network for autonomous driving[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C, USA:IEEE Press, 2017:1907-1915.
[13] SU H, MAJI S, KALOGERAKIS E, et al.Multi-view convolutional neural networks for 3D shape recognition[C]//Proceedings of 2017 IEEE International Conference on Computer Vision.Washington D.C, USA:IEEE Press, 2017:945-953.
[14] FENG Y, ZHANG Z, ZHAO X, et al.GVCNN:group-view convolutional neural networks for 3d shape recognition[C]//Proceedings of 2018 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C, USA:IEEE Press, 2018:264-272.
[15] MATURANA D, SCHERER S.VoxNet:a 3D convolutional neural network for real-time object recognition[C]//Proceedings of 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems.Washington D.C, USA:IEEE Press, 2018:922-928.
[16] CHARLES R Q, HAO S, MO K C, et al.PointNet:deep learning on point sets for 3D classification and segmentation[C]//Proceedings of 2017 IEEE Conference on Computer Vision and Pattern Recognition.Washington D.C, USA:IEEE Press, 2017:652-660.
[17] ZHAO H S, JIANG L, JIA J Y, et al.Point transformer[EB/OL].[2021-10-11].https://arxiv.org/abs/2012.09164.
[18] 朱威, 绳荣金, 汤如, 等.基于动态图卷积和空间金字塔池化的点云深度学习网络[J].计算机科学, 2020, 47(7):192-198. ZHU W, SHENG R J, TANG R, et al.Point cloud deep learning network based on dynamic graph convolution and spatial pyramid pooling[J].Computer Science, 2020, 47(7):192-198.(in Chinese)
[19] LI Y, BU R, SUN M, et al.PointCNN:convolution on x-transformed points[J].Advances in Neural Information Processing Systems, 2018, 31:820-830.
[20] 顾砾, 季怡, 刘纯平.基于多模态特征融合的三维点云分类方法[J].计算机工程, 2021, 47(2):279-284. GU L, JI Y, LIU C P.Classification method of three-dimensional point cloud based on multiple modal feature fusion[J].Computer Engineering, 2021, 47(2):279-284.(in Chinese)
[21] 田钰杰, 管有庆, 龚锐.一种鲁棒的多特征点云分类分割深度神经网络[J].计算机工程, 2021, 47(11):234-240. TIAN Y J, GUAN Y Q, GONG R.A robust deep neural network for multi-feature point cloud classification and segmentation[J].Computer Engineering, 2021, 47(11):234-240.(in Chinese)
[22] VASWANI A, SHAZEER N, PARMAR N, et al.Attention is all you need[C]//Proceedings of 2017 IEEE Conference on Neural Information Processing Systems.Washington D.C., USA:IEEE Press, 2017:5998-6008.
[23] HAN Z Z, LU H L, LIU Z B, et al.3D2SeqViews:aggregating sequential views for 3D global feature learning by CNN with hierarchical attention aggregation[J].IEEE Transactions on Image Processing, 2019, 28(8):3986-3999.
[24] HAN Z Z, SHANG M Y, LIU Z B, et al.SeqViews2SeqLabels:learning 3D global features via aggregating sequential views by RNN with attention[J].IEEE Transactions on Image Processing, 2019, 28(2):658-672.
[25] LIU X, HAN Z, LIU Y S, et al.Point2sequence:learning the shape representation of 3D point clouds with an attention-based sequence to sequence network[C]//Proceedings of 2019 AAAI Conference on Artificial Intelligence.Palo Alto, USA:AAAI Press, 2019:8778-8785.

选择文件类型/文献管理软件名称

选择包含的内容