arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.13135cs.CVcs.LGeess.IV
收藏

Permutation Matters:用于点云学习的各向异性卷积层

Permutation Matters: Anisotropic Convolutional Layer for Learning on Point Clouds

Artificial Intelligence Institute(人工智能研究所);Shanghai Jiao Tong University(上海交通大学)

Zhongpai Gao, Guangtao Zhai, Junchi Yan, Xiaokang Yang

摘要:

在许多3D计算机视觉应用中,对点云的有效表示学习的需求日益增长。卷积神经网络(CNN)成功的背后是数据(例如,图像)是欧几里得结构的。然而,点云是不规则和无序的。各种点神经网络已经开发出各向同性滤波器或使用加权矩阵来克服点云上的结构不一致性。然而,各向同性滤波器或加权矩阵限制了表示能力。在本文中,我们提出了一个可置换的各向异性卷积运算(PAI-Conv),根据球体表面上均匀分布的核点集,使用点积注意力计算每个点的软置换矩阵,并执行共享各向异性滤波器。实际上,具有核点的点积与自然语言处理(NLP)中广泛使用的Transformer中具有键的点积类似。从这个角度来看,PAI-Conv可以被视为点云的Transformer,它具有物理意义,并且与高效的随机点采样方法配合使用。点云上的综合实验表明,与最先进的方法相比,PAI-Conv在分类和语义分割任务中产生了有竞争力的结果。

英文摘要:

It has witnessed a growing demand for efficient representation learning on point clouds in many 3D computer vision applications. Behind the success story of convolutional neural networks (CNNs) is that the data (e.g., images) are Euclidean structured. However, point clouds are irregular and unordered. Various point neural networks have been developed with isotropic filters or using weighting matrices to overcome the structure inconsistency on point clouds. However, isotropic filters or weighting matrices limit the representation power. In this paper, we propose a permutable anisotropic convolutional operation (PAI-Conv) that calculates soft-permutation matrices for each point using dot-product attention according to a set of evenly distributed kernel points on a sphere's surface and performs shared anisotropic filters. In fact, dot product with kernel points is by analogy with the dot-product with keys in Transformer as widely used in natural language processing (NLP). From this perspective, PAI-Conv can be regarded as the transformer for point clouds, which is physically meaningful and is robust to cooperate with the efficient random point sampling method. Comprehensive experiments on point clouds demonstrate that PAI-Conv produces competitive results in classification and semantic segmentation tasks compared to state-of-the-art methods.

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。

↑