[论文解读] RGCNN: Regularized Graph CNN for Point Cloud Segmentation

Gusi Te, Wei Hu+2 more|arXiv (Cornell University)|Jun 8, 2018

3D Shape Modeling and Analysis参考文献 23被引用 90

一句话总结

RGCNN 直接处理不规则的3D点云，使用动态更新的图拉普拉斯算子和图信号平滑先验，在分割和分类中获得具竞争力的精度并提高鲁棒性。

ABSTRACT

Point cloud, an efficient 3D object representation, has become popular with the development of depth sensing and 3D laser scanning techniques. It has attracted attention in various applications such as 3D tele-presence, navigation for unmanned vehicles and heritage reconstruction. The understanding of point clouds, such as point cloud segmentation, is crucial in exploiting the informative value of point clouds for such applications. Due to the irregularity of the data format, previous deep learning works often convert point clouds to regular 3D voxel grids or collections of images before feeding them into neural networks, which leads to voluminous data and quantization artifacts. In this paper, we instead propose a regularized graph convolutional neural network (RGCNN) that directly consumes point clouds. Leveraging on spectral graph theory, we treat features of points in a point cloud as signals on graph, and define the convolution over graph by Chebyshev polynomial approximation. In particular, we update the graph Laplacian matrix that describes the connectivity of features in each layer according to the corresponding learned features, which adaptively captures the structure of dynamic graphs. Further, we deploy a graph-signal smoothness prior in the loss function, thus regularizing the learning process. Experimental results on the ShapeNet part dataset show that the proposed approach significantly reduces the computational complexity while achieving competitive performance with the state of the art. Also, experiments show RGCNN is much more robust to both noise and point cloud density in comparison with other methods. We further apply RGCNN to point cloud classification and achieve competitive results on ModelNet40 dataset.

研究动机与目标

• 在不进行体素化或渲染的情况下，推动对不规则三维点云的直接学习。
• 提出一种正则化的图卷积神经网络，在各层更新图结构以捕捉动态拓扑。
• 在损失中引入图信号平滑先验，以实现拉普拉斯平滑。
• 展示在计算复杂度降低的同时具有竞争力的分割性能，并对噪声/密度的鲁棒性有所提升。
• 将该架构扩展到 ModelNet40 的点云分类。

提出的方法

• 使用坐标和法向量作为节点特征，将点云表示为图。
• 构建全连接图，边权为 a_{i,j}=exp(-β||p_i-p_j||^2)，以捕捉全局关系。
• 使用带有 Chebyshev 多项式近似的谱图卷积进行局部化过滤。
• 在每一层更新图拉普拉斯矩阵，以适应随特征驱动的拓扑演变。
• 将图信号平滑先验 y^T L y 纳入损失，以在各层之间促进拉普拉斯平滑。
• 给出置换不变性证明，确保输出与输入点的顺序对齐。

实验结果

研究问题

• RQ1 图-CNN 是否能够在不进行体素化的情况下直接对不规则点云进行处理？
• RQ2 逐层更新图拉普拉斯是否提升分割性能和鲁棒性？
• RQ3 通过强化拓扑一致性，图信号平滑先验是否能改善分割？
• RQ4 在 ShapeNet Part 的分割和 ModelNet40 的分类方面，RGCNN 与最先进方法相比如何？

主要发现

• RGCNN 在 ShapeNet Part 的平均 IoU 与 PointNet、PointNet++ 和 SynSpecCNN 相当。
• 全连接图构造的性能优于 k-NN 图（80.4% 的平均 IoU，相对于 k-NN 的较低值）。
• RGCNN 对噪声和点密度降低具有鲁棒性，在扰动和缺失数据下超越基线。
• 在 ModelNet40 上，该方法的分类准确率与 PointNet 与 PointNet++ 相当，尽管略低于最佳的当代方法。
• 该架构在时空复杂度方面表现良好，前向时间快、模型规模可扩展。