Spherical data is found in many applications. By modeling the discretized sphere as a graph, we can accommodate non-uniformly distributed, partial, and changing samplings. Moreover, graph convolutions are computationally more efficient than spherical convolutions. As equivariance is desired to exploit rotational symmetries, we discuss how to approach rotation equivariance using the graph neural network introduced in Defferrard et al. (2016). Experiments show good performance on rotation-invariant learning problems. Code and examples are available at https://github.com/SwissDataScienceCenter/DeepSphere
@article{arxiv.1904.05146,
title = {DeepSphere: towards an equivariant graph-based spherical CNN},
author = {Michaël Defferrard and Nathanaël Perraudin and Tomasz Kacprzak and Raphael Sgier},
journal= {arXiv preprint arXiv:1904.05146},
year = {2019}
}
Comments
published at the ICLR 2019 Workshop on Representation Learning on Graphs and Manifolds. arXiv admin note: text overlap with arXiv:1810.12186