Self-Supervised Point Cloud Completion via Inpainting
Self-Supervised Point Cloud Completion via Inpainting
复制标题
DOI：
--
复制
发表时间：
2021-11
期刊：
影响因子：
--
通讯作者：
Himangi Mittal;Brian Okorn;Arpit Jangid;David Held
中科院分区：
文献类型：
--
作者：
Himangi Mittal;Brian Okorn;Arpit Jangid;David Held
在城市环境中导航时，许多需要跟踪和避让的物体都被严重遮挡。利用这些局部扫描进行规划和跟踪可能具有挑战性。这项工作的目的是学习补全这些局部点云，仅通过局部观测就能全面了解物体的几何形状。先前的方法是在目标物体完整的真实标注的帮助下实现这一点的，而这些标注仅适用于模拟数据集。然而，对于现实世界的激光雷达数据，这样的真实标注是不可用的。在这项工作中，我们提出了一种自监督点云补全算法——PointPnCNet，它仅在局部扫描上进行训练，而不假定能够获取完整的真实标注。我们的方法通过修复来实现这一点。我们去除一部分输入数据，并训练网络补全缺失的区域。由于很难确定初始点云中哪些区域被遮挡，哪些是人为去除的，我们的网络学习补全整个点云，包括初始局部点云中缺失的区域。我们表明，我们的方法在合成数据集ShapeNet和现实世界的激光雷达数据集Semantic KITTI上都优于先前的无监督和弱监督方法。
When navigating in urban environments, many of the objects that need to be tracked and avoided are heavily occluded. Planning and tracking using these partial scans can be challenging. The aim of this work is to learn to complete these partial point clouds, giving us a full understanding of the object's geometry using only partial observations. Previous methods achieve this with the help of complete, ground-truth annotations of the target objects, which are available only for simulated datasets. However, such ground truth is unavailable for real-world LiDAR data. In this work, we present a self-supervised point cloud completion algorithm, PointPnCNet, which is trained only on partial scans without assuming access to complete, ground-truth annotations. Our method achieves this via inpainting. We remove a portion of the input data and train the network to complete the missing region. As it is difficult to determine which regions were occluded in the initial cloud and which were synthetically removed, our network learns to complete the full cloud, including the missing regions in the initial partial cloud. We show that our method outperforms previous unsupervised and weakly-supervised methods on both the synthetic dataset, ShapeNet, and real-world LiDAR dataset, Semantic KITTI.