Research map / Architectures and efficiency / CNN design
Spherical CNNs on Unstructured Grids
Read the paperCheck 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 parameterized differential operators instead of conventional mesh convolution kernels | Spherical CNN for MNIST · Spherical MNIST | accuracy (%): 96.00 -> 99.23 | Helped |
| Use parameterized differential operators instead of conventional mesh convolution kernels | Spherical CNN for MNIST · Spherical MNIST | accuracy (%): 94.41 -> 99.23 | Helped |
| Use the full combination of differential operators in the kernel (ablation) | MeshConv · ModelNet40 | accuracy | Helped |
| Process the entire global input rather than random crops | Climate segmentation model · Climate pattern segmentation | mean accuracy (%): 78.67 -> 94.67 | Helped |
| Use spherical MeshConv segmentation instead of planar UNet | 2D3DS semantic segmentation · 2D3DS | mean per-class accuracy: 0.5080 -> 0.5465 | Helped |
| Use spherical MeshConv segmentation instead of planar FCN8s | 2D3DS semantic segmentation · 2D3DS | Mean IoU: 0.3560 -> 0.3829 | Helped |
| Use spherical MeshConv segmentation instead of PointNet++ | 2D3DS semantic segmentation · 2D3DS | mean per-class accuracy: 0.3349 -> 0.5465 | Helped |
| Use spherical MeshConv segmentation instead of PointNet++ | 2D3DS semantic segmentation · 2D3DS | Mean IoU: 0.2312 -> 0.3829 | Helped |
| Use orientation-aware features rather than rotational invariance | Spherical CNN · Spherical MNIST | accuracy (%): 96.00 -> 99.23 | Helped |
Closest papers by meaning
- Scaling Spherical CNNs (ICML 2023 · similarity 0.88)
- PDO-eS2CNNs: Partial Differential Operator Based Equivariant Spherical CNNs (AAAI 2021 · similarity 0.87)
- Scalable and Equivariant Spherical CNNs by Discrete-Continuous (DISCO) Convolutions (ICLR 2023 · similarity 0.87)
- Spherical CNNs (ICLR 2018 · similarity 0.87)
- Neural Operators with Localized Integral and Differential Kernels (ICML 2024 · similarity 0.84)
- Attention on the Sphere (NeurIPS 2025 · similarity 0.84)
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.