Research map / Graph learning / GNN architectures
Rotation Equivariant Graph Convolutional Network for Spherical Image Classification
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use GICOPix rather than GSSPix for spherical graph construction | SGCN · S-MNIST N/R | accuracy (%): 43.26 -> 93.43 | Helped |
| Use GICOPix rather than HEALPix for spherical graph construction | SGCN · S-MNIST N/R | accuracy (%): 91.41 -> 93.43 | Helped |
| Train on non-rotated examples and test on rotated examples with GICOPix-SGCN instead of SphereNet | S-MNIST · S-MNIST N/R | accuracy (%): 55.18 -> 93.43 | Helped |
| Use GICOPix rather than HEALPix for spherical graph construction | SGCN · S-CIFAR-10 R/R | accuracy (%): 55.08 -> 58.03 | Helped |
| Use GICOPix rather than HEALPix for spherical graph construction | SGCN · S-CIFAR-10 N/R | accuracy (%): 51.90 -> 56.84 | Helped |
| Use GICOPix-SGCN instead of SphereNet for non-rotated training and rotated testing | S-CIFAR-10 · S-CIFAR-10 N/R | accuracy (%): 37.18 -> 56.84 | Helped |
| Use GICOPix rather than HEALPix for spherical graph construction | SGCN · ModelNet40 N/R | accuracy (%): 78.4 -> 84.0 | Helped |
| Use GICOPix-SGCN instead of PDOs for non-rotated training and rotated testing | ModelNet40 · ModelNet40 N/R | accuracy (%): 23.5 -> 84.0 | Helped |
| Add the proposed hierarchical pooling instead of common pooling (ablation) | GICOPix-SGCN · S-MNIST N/R | accuracy (%): 92.67 -> 93.43 | Helped |
| Add the proposed hierarchical pooling instead of common pooling (ablation) | GICOPix-SGCN · ModelNet40 N/R | accuracy (%): 83.31 -> 84.04 | Helped |
| Add the proposed transition layer instead of global average pooling (ablation) | GICOPix-SGCN · S-MNIST N/R | accuracy (%): 91.85 -> 93.43 | Helped |
| Add the proposed transition layer instead of global average pooling (ablation) | GICOPix-SGCN · ModelNet40 N/R | accuracy (%): 81.97 -> 84.04 | Helped |
Closest papers by meaning
- Spin-Weighted Spherical CNNs (NeurIPS 2020 · similarity 0.91)
- DeepSphere: a graph-based spherical CNN (ICLR 2020 · similarity 0.91)
- Efficient Generalized Spherical CNNs (ICLR 2021 · similarity 0.90)
- Equivariance versus Augmentation for Spherical Images (ICML 2022 · similarity 0.90)
- Scalable and Equivariant Spherical CNNs by Discrete-Continuous (DISCO) Convolutions (ICLR 2023 · similarity 0.90)
- Cross-Domain 3D Equivariant Image Embeddings (ICML 2019 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.