PyTorch implementation of the spherical-graph-CNN described in DeepSphere: a Graph-Based Spherical CNN, Defferrard et al. (2019) [1] and reproduction of the experiments from the paper.
This is a graph-based spherical convolutional neural network that strikes an interesting balance of trade-offs for a wide variety of applications.
Abstract:
Designing a convolution for a spherical neural network requires a delicate tradeoff between efficiency and rotation equivariance. DeepSphere, a method based on a graph representation of the discretized sphere, strikes a controllable balance between these two desiderata. This contribution is twofold. First, we study both theoretically and empirically how equivariance is affected by the underlying graph with respect to the number of pixels and neighbors. Second, we evaluate DeepSphere on relevant problems. Experiments show state-of-the-art performance and demonstrates the efficiency and flexibility of this formulation. Perhaps surprisingly, comparison with previous work suggests that anisotropic filters might be an unnecessary price to pay.
[1] DeepSphere: a Graph-Based Spherical CNN, Michaël Defferrard, Martino Milani, Frédérick Gusset and Nathanaël Perraudin, 2019 (ICLR 2020)