Tranquil Clouds: Neural Networks for Learning Temporally Coherent Features in Point Clouds
arXiv:1907.05279
Abstract
Point clouds, as a form of Lagrangian representation, allow for powerful and flexible applications in a large number of computational disciplines. We propose a novel deep-learning method to learn stable and temporally coherent feature spaces for points clouds that change over time. We identify a set of inherent problems with these approaches: without knowledge of the time dimension, the inferred solutions can exhibit strong flickering, and easy solutions to suppress this flickering can result in undesirable local minima that manifest themselves as halo structures. We propose a novel temporal loss function that takes into account higher time derivatives of the point positions, and encourages mingling, i.e., to prevent the aforementioned halos. We combine these techniques in a super-resolution method with a truncation approach to flexibly adapt the size of the generated positions. We show that our method works for large, deforming point sets from different sources to demonstrate the flexibility of our approach.
arXiv commentFurther information and videos at https://ge.in.tum.de/publications/2020-iclr-prantl/
References in corpus (9)
- Tensor field networks: Rotation- and translation-equivariant neural networks for 3D point clouds
- PCPNET: Learning Local Shape Properties from Raw Point Clouds
- Learning Particle Dynamics for Manipulating Rigid Bodies, Deformable Objects, and Fluids
- PointConv: Deep Convolutional Networks on 3D Point Clouds
- Complex-YOLO: Real-time 3D Object Detection on Point Clouds
- A Point Set Generation Network for 3D Object Reconstruction from a Single Image
- Spherical Convolutional Neural Network for 3D Point Clouds
- PointGrow: Autoregressively Learned Point Cloud Generation with Self-Attention
- URSA: A Neural Network for Unordered Point Clouds Using Constellations