Research map / 3D vision / Face and shape
Mesh-Based Autoencoders for Localized Deformation Component Analysis
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 nonlinear autoencoder reconstruction instead of linear combination of extracted components | Mesh-based autoencoder · Swing unseen data, 50 components | E_rms: 25.2994 -> 14.0836 | Helped |
| Use nonlinear autoencoder reconstruction instead of linear combination of extracted components | Mesh-based autoencoder · Swing unseen data, 50 components | STED: 0.05214 -> 0.03789 | Helped |
| Use nonlinear autoencoder reconstruction instead of linear combination of extracted components | Mesh-based autoencoder · SCAPE unseen data, 50 components | E_rms: 17.1754 -> 13.556 | Helped |
| Set λ₁ and λ₂ to 0.4 instead of 0.5 (ablation) | Mesh-based autoencoder · SCAPE unseen data, 50 components | E_rms: 13.556 -> 14.1763 | No clear effect |
| Set λ₁ and λ₂ to 0.6 instead of 0.5 (ablation) | Mesh-based autoencoder · SCAPE unseen data, 50 components | E_rms: 13.556 -> 13.6429 | No clear effect |
| Set λ₁ to 0.4 and λ₂ to 0.6 instead of 0.5 (ablation) | Mesh-based autoencoder · SCAPE unseen data, 50 components | E_rms: 13.556 -> 14.2645 | No clear effect |
| Set λ₁ to 0.6 and λ₂ to 0.4 instead of 0.5 (ablation) | Mesh-based autoencoder · SCAPE unseen data, 50 components | E_rms: 13.556 -> 14.093 | No clear effect |
| Replace tanh with ReLU activations (ablation) | Mesh-based autoencoder · Quantitative analysis | reconstruction error | Hurt |
| Use nonlinear autoencoder reconstruction rather than linear component combinations | Mesh-based autoencoder · SCAPE unseen data, 50 components | E_rms: 17.1754 -> 13.556 | Helped |
Closest papers by meaning
- Variational Autoencoders for Deforming 3D Mesh Models (CVPR 2018 · similarity 0.91)
- Fully Convolutional Mesh Autoencoder using Efficient Spatially Varying Kernels (NeurIPS 2020 · similarity 0.90)
- 4Deform: Neural Surface Deformation for Robust Shape Interpolation (CVPR 2025 · similarity 0.86)
- LaGeM: A Large Geometry Model for 3D Representation Learning and Diffusion (ICLR 2025 · similarity 0.85)
- Sparc3D: Sparse Representation and Construction for High-Resolution 3D Shapes Modeling (NeurIPS 2025 · similarity 0.85)
- SPU-PMD: Self-Supervised Point Cloud Upsampling via Progressive Mesh Deformation (CVPR 2024 · similarity 0.85)
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.