Differentiable Iterative Surface Normal Estimation
Preprint
Abstract
This paper presents an end-to-end differentiable algorithm for anisotropic surface normal estimation on unstructured point-clouds. We utilize graph neural networks to iteratively infer point weights for a plane fitting algorithm applied to local neighborhoods. The approach retains the interpretability and efficiency of traditional sequential plane fitting while benefiting from a data-dependent deep-learning parameterization. This results in a state-of-the-art surface normal estimator that is robust to noise, outliers and point density variation and that preserves sharp features through anisotropic kernels and a local spatial transformer. Contrary to previous deep learning methods, the proposed approach does not require any hand-crafted features while being faster and more parameter efficient.Keywords
All Related Versions
This publication has 14 references indexed in Scilit:
- PointCleanNet: Learning to Denoise and Remove Outliers from Dense Point CloudsComputer Graphics Forum, 2019
- Nesti-Net: Normal Estimation for Unstructured 3D Point Clouds Using Convolutional Neural Networks2019
- PCPNet Learning Local Shape Properties from Raw Point CloudsComputer Graphics Forum, 2018
- Geometric Deep Learning: Going beyond Euclidean dataIEEE Signal Processing Magazine, 2017
- Deep Learning for Robust Normal Estimation in Unstructured Point CloudsComputer Graphics Forum, 2016
- Discriminatively Trained Dense Surface Normal Estimation2014
- Fast and Robust Normal Estimation for Point Clouds with Sharp FeaturesComputer Graphics Forum, 2012
- Voronoi-Based Curvature and Feature Estimation from Point CloudsIEEE Transactions on Visualization and Computer Graphics, 2010
- Multiple View Geometry in Computer Vision2004
- The approximation power of moving least-squaresMathematics of Computation, 1998