Normal estimation for unstructured point clouds is an important task in 3D computer vision. Current methods achieve encouraging results by mapping local patches to normal vectors or learning local surface fitting using neural networks. However, these methods are not generalized well to unseen scenarios and sensitive to parameter settings. To resolve these issues, we propose an implicit function to learn an angle field around the normal of each point in the spherical coordinate system, which is dubbed as Neural Angle Fields (NeAF). Rather than directly predicting the normal of an input point, we predict the angle offset between the ground truth normal and a randomly sampled query normal. This strategy pushes the network to observe more diverse samples, which leads to higher prediction accuracy in a more robust manner. To predict normals from the learned angle fields at inference time, we randomly sample query vectors in a unit spherical space, and take the vectors with minimal angle values as the predicted normals. To further leverage the prior learned by NeAF, we propose to refine the predicted normal vectors by minimizing the angle offsets. The experimental results with synthetic and real scanned data show significant improvements over the state-of-the-art under widely used benchmarks.
Overview of our method. The NeAF is designed to estimate normals for point clouds by learning implicit angle fields. Given a query vector sampled on the unit sphere and a local patch as input, the network outputs the angle offset between the query vector and the ground truth normal of the patch. After convergence, an angle field centered on the ground truth normal is formed. At inference time, we use a coarse-to-fine strategy to predict the final normal with an angle offset zero from the learned angle field.
@inproceedings{li2023NeAF,
title={NeAF: Learning Neural Angle Fields for Point Normal Estimation},
author={Li, Shujuan and Zhou, Junsheng and Ma, Baorui and Liu, Yu-Shen and Han, Zhizhong},
booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
year={2023}
}