arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2603.04099cs.CVcs.AI
收藏

高效点云处理与高维位置编码及非局部MLP

Efficient Point Cloud Processing with High-Dimensional Positional Encoding and Non-Local MLPs

School of Artificial Intelligence and Robotics, Quanzhou Institute of Industrial Design and Machine Intelligence Innovation, Hunan University(人工智能与机器人学院、泉州工业设计与智能机械创新研究院、湖南大学);College of Engineering and Design, Hunan Normal University(工程与设计学院、湖南师范大学);College of Intelligence Science and Technology, National University of Defense Technology(智能科学与技术学院、国防科技大学);School of Computing and Information Systems, The University of Melbourne(计算与信息学院、墨尔本大学)

Yanmei Zou, Hongshan Yu, Yaonan Wang, Zhengeng Yang, Xieyuanli Chen, Kailun Yang, Naveed Akhtar

AI总结:

本文提出了一种基于高维位置编码和非局部MLP的点云处理方法,通过两阶段抽象和细化框架提升效率与效果。

摘要:

多层感知机(MLP)模型是当前点云处理的基础。然而,其复杂的网络架构模糊了其优势的来源,并限制了这些模型的应用。在本文中,我们开发了一个两阶段抽象和细化(ABS-REF)视图,用于点云处理中的模块化特征提取。该视图阐明了尽管早期模型专注于ABS阶段,但更近期的技术则设计了复杂的REF阶段以获得性能优势。然后,我们提出了一种高维位置编码(HPE)模块,以显式利用内在位置信息,将“位置编码”概念从Transformer文献中扩展开来。HPE可以轻松部署在基于MLP的架构中,并且与基于Transformer的方法兼容。在我们的ABS-REF视图中,我们重新思考了基于MLP方法中的局部聚合,并提出用非局部MLP代替耗时的局部MLP操作,用于捕捉邻居之间的局部关系。相反,我们使用非局部MLP进行高效的非局部信息更新,并结合所提出的HPE以实现有效的局部信息表示。我们利用这些模块开发了HPENets,一组遵循ABS-REF范式的MLP网络,集成了可扩展的基于HPE的REF阶段。在七个公共数据集上进行的广泛实验表明,HPENets在效率和有效性之间实现了良好的平衡。值得注意的是,HPENet在ScanObjectNN、S3DIS、ScanNet和ShapeNetPart上分别仅使用50.0%、21.5%、23.1%和44.4%的FLOPs,就超过了PointNeXt这一强大的基于MLP的对手,分别在mAcc、mIoU、mIoU和Cls.mIoU上超过了1.1%、4.0%、1.8%和0.2%。源代码可在https://github.com/zouyanmei/HPENet_v2.git上获得。

英文摘要:

Multi-Layer Perceptron (MLP) models are the foundation of contemporary point cloud processing. However, their complex network architectures obscure the source of their strength and limit the application of these models. In this article, we develop a two-stage abstraction and refinement (ABS-REF) view for modular feature extraction in point cloud processing. This view elucidates that whereas the early models focused on ABS stages, the more recent techniques devise sophisticated REF stages to attain performance advantages. Then, we propose a High-dimensional Positional Encoding (HPE) module to explicitly utilize intrinsic positional information, extending the ``positional encoding'' concept from Transformer literature. HPE can be readily deployed in MLP-based architectures and is compatible with transformer-based methods. Within our ABS-REF view, we rethink local aggregation in MLP-based methods and propose replacing time-consuming local MLP operations, which are used to capture local relationships among neighbors. Instead, we use non-local MLPs for efficient non-local information updates, combined with the proposed HPE for effective local information representation. We leverage our modules to develop HPENets, a suite of MLP networks that follow the ABS-REF paradigm, incorporating a scalable HPE-based REF stage. Extensive experiments on seven public datasets across four different tasks show that HPENets deliver a strong balance between efficiency and effectiveness. Notably, HPENet surpasses PointNeXt, a strong MLP-based counterpart, by 1.1% mAcc, 4.0% mIoU, 1.8% mIoU, and 0.2% Cls. mIoU, with only 50.0%, 21.5%, 23.1%, 44.4% of FLOPs on ScanObjectNN, S3DIS, ScanNet, and ShapeNetPart, respectively. Source code is available at https://github.com/zouyanmei/HPENet_v2.git.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑