Raw point clouds obtained from real-world scanning are always incomplete and ununiformly distributed, which would result in structural losses of object shapes and bring about difficulties in further high-level 3D vision tasks. Therefore, a learning-based method called CRA-Net is proposed in this paper to repair partial point clouds and predict complete object shapes. Compared with most existing networks that only leverage global features, CRA-Net successfully utilizes local features to restore clearer details of object shapes with low instability. First, we propose an adaptive neighborhood query method that is able to adjust query centers and radiuses to cover different object shapes and acquire balanced local regions. Second, we build a parallel encoder to extract multiscale features from the input. Third, we design a cross-regional attention module based on graph attention network. It quantifies underlying relationships among all the local features under certain conditions interpreted by global features. Based on such relationships, each conditional local feature vector is able to search across the regions and selectively absorb other local features. Fourth, we design a coarse decoder to collect these cross region features and generate the skeleton of complete point cloud. Finally, we refine the coarse point cloud by comparing it with the input, and up sample it using folding-based layers.

Our network is first trained and tested on manually made partial-complete point clouds pairs generated by the scanning process of a virtual LiDAR on eight categories of objects. Then it is tested on real-world point clouds of indoor and outdoor scenes. Compared with existing representative methods, our CRA-Net always restores the most accurate point clouds with the clearest details.

An adaptive neighborhood query algorithm that acquires balanced local regions for partial point cloud. A cross-regional attention module that learns relationships among local regions under shared global conditions. A two-step decoder that generates complete point clouds based on cross-region features from coarse to fine. Better performances on both manually-made and real-world datasets, and a more comprehensive relation between local areas of inputs and outputs.

从实际扫描中获得的原始点云始终不完整且分布不均匀，这将导致对象形状的结构损失，并给进一步的高级3D视觉任务带来困难。因此，本文提出了一种基于学习的方法，称为CRA-Net，用于修复部分点云并预测完整的对象形状。与大多数仅利用全局功能的现有网络相比，CRA-Net成功地利用了本地功能，以较低的不稳定性恢复了对象形状的更清晰的细节。首先，我们提出了一种自适应邻域查询方法，该方法能够调整查询中心和半径以覆盖不同的对象形状并获取平衡的局部区域。其次，我们构建一个并行编码器以从输入中提取多尺度特征。第三，我们设计了一个基于图注意力网络的跨区域注意力模块。它在全局条件解释的特定条件下，量化了所有局部特征之间的潜在关系。基于这样的关系，每个条件局部特征向量都能够跨区域搜索并有选择地吸收其他局部特征。第四，我们设计了一个粗略的解码器来收集这些跨区域特征并生成完整点云的骨架。最后，我们通过将粗点云与输入进行比较来细化粗点云，并使用基于折叠的图层对其进行上采样。每个条件局部特征向量都能够搜索区域并有选择地吸收其他局部特征。第四，我们设计了一个粗略的解码器来收集这些跨区域特征并生成完整点云的骨架。最后，我们通过将粗点云与输入进行比较来细化粗点云，并使用基于折叠的图层对其进行上采样。每个条件局部特征向量都能够搜索区域并有选择地吸收其他局部特征。第四，我们设计了一个粗略的解码器来收集这些跨区域特征并生成完整点云的骨架。最后，我们通过将粗点云与输入进行比较来细化粗点云，并使用基于折叠的图层对其进行上采样。

我们的网络首先经过人工制作的部分完成点云对的训练和测试，这些点对是通过对八个类别的对象进行虚拟LiDAR扫描过程生成的。然后在室内和室外场景的真实点云上进行测试。与现有的代表性方法相比，我们的CRA-Net始终以最清晰的细节还原最准确的点云。

一种自适应的邻域查询算法，该算法为局部点云获取平衡的局部区域。跨区域关注模块，用于在共享的全球条件下学习局部区域之间的关系。两步解码器可根据从粗糙到精细的跨区域特征生成完整的点云。在手工制作的数据集和真实的数据集上都有更好的性能，并且输入和输出的局部区域之间的关系更加全面。