Research map / Architectures and efficiency / CNN design
Harmonic Networks: Deep Translation and Rotation Equivariance
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 circular-harmonic filters with rotation-equivariant feature streams instead of regular CNN filters | Rotated-MNIST classifier · Rotated MNIST test | test error (%): 5.03 -> 1.69 | Helped |
| Use circular-harmonic filters with rotation-equivariant feature streams instead of regular CNN filters | Rotated-MNIST classifier · Rotated MNIST test | test error (%): 3.50 -> 1.69 | Helped |
| Use circular-harmonic filters with rotation-equivariant feature streams instead of P4CNN | Rotated-MNIST classifier · Rotated MNIST test | test error (%): 2.28 -> 1.69 | Helped |
| Hard-bake local rotation equivariance into the network | Boundary detector · BSD500 | ODS: 0.702 -> 0.726 | Helped |
| Hard-bake local rotation equivariance into the network | Boundary detector · BSD500 | OIS: 0.715 -> 0.742 | Helped |
| Add a sparsity regularizer to the H-DSN objective | H-DSN · BSD500 | boundary detection performance | Helped |
| Use hard-baked rotation equivariance rather than learning rotation structure from more training data | H-Net versus regular CNN · Rotated MNIST, varying training-set sizes | test accuracy | Helped |
Closest papers by meaning
- Empowering Networks With Scale and Rotation Equivariance Using A Similarity Convolution (ICLR 2023 · similarity 0.91)
- Exploiting Cyclic Symmetry in Convolutional Neural Networks (ICML 2016 · similarity 0.90)
- Polar Transformer Networks (ICLR 2018 · similarity 0.89)
- Rotation-Equivariant Self-Supervised Method in Image Denoising (CVPR 2025 · similarity 0.87)
- On Making $SE(2)$-Invariant Networks Optimal (NeurIPS 2026 · similarity 0.87)
- A Probabilistic Approach to Learning the Degree of Equivariance in Steerable CNNs (ICML 2024 · similarity 0.86)
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.