Research map / Probabilistic and scientific modeling / Molecules and proteins
Parametrizing Product Shape Manifolds by Composite Networks
Josua Sassen, Klaus Hildebrandt, Martin Rumpf, Benedikt Wirth
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 a composite factorwise MLP plus convolutional combination network instead of a monolithic network | Freaky Torus parametrization · Freaky Torus | R²: 0.95 -> 0.99 | Improved |
| Use a composite network instead of affine combination | SPGA shape manifolds · SMPL+X | R²: 0.78 -> 0.93 | Improved |
| Use a composite network instead of a monolithic network | SPGA shape manifolds · SMPL+X | R²: 0.85 -> 0.93 | Improved |
| Use a composite network instead of affine combination | SPGA shape manifolds · SCAPE | R²: 0.77 -> 0.91 | Improved |
| Use a composite network instead of a monolithic network | SPGA shape manifolds · SCAPE | R²: 0.60 -> 0.91 | Improved |
| Use a composite network instead of affine combination | SPGA shape manifolds · Hands | R²: 0.88 -> 0.98 | Improved |
| Use a composite network instead of a monolithic network | SPGA shape manifolds · Hands | R²: 0.95 -> 0.98 | Improved |
| Use a composite network instead of affine combination | SPGA shape manifolds · Faces | R²: 0.96 -> 0.99 | Improved |
| Use a composite network instead of a monolithic network | SPGA shape manifolds · Faces | R²: 0.95 -> 0.99 | Improved |
| Use a composite network instead of the prior method | SPGA shape manifolds · Hands | R²: 0.80 -> 0.98 | Improved |
| Use a composite network instead of the prior method | SPGA shape manifolds · Faces | R²: 0.95 -> 0.99 | Improved |
| Reduce training data to 20% of the full training set | Composite network · SCAPE | R²: 0.91 -> 0.86 | Hurt |
Closest papers by meaning
- Representation Learning via Manifold Flattening and Reconstruction (JMLR 2024 · similarity 0.88)
- Efficient Representation of Low-Dimensional Manifolds using Deep Networks (ICLR 2017 · similarity 0.88)
- Learning Manifold Patch-Based Representations of Man-Made Shapes (ICLR 2021 · similarity 0.88)
- Learning Eigenstructures of Unstructured Data Manifolds (CVPR 2026 · similarity 0.88)
- Building Transformation Layers for Riemannian Neural Networks (NeurIPS 2026 · similarity 0.87)
- Matrix Manifold Neural Networks++ (ICLR 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.
