Research map / 3D vision / Face and shape
Adaptive Spiral Layers for Efficient 3D Representation Learning on Meshes
Francesca Babiloni, Matteo Maggioni, Thomas Tanay, Jiankang Deng, Ales Leonardis, Stefanos Zafeiriou
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 |
|---|---|---|---|
| replace static SpiralConv layers with adaptive spiral layers (FAUST) | SpiralNet++ · FAUST correspondence | parameters (M): 1.91 -> 0.31 at same 99.88 accuracy | Improved |
| replace static SpiralConv layers with adaptive spiral layers (FAUST) | SpiralNet++ · FAUST correspondence | GFlops: 13.04 -> 0.82 | Improved |
| add adaptive spiral operator in 3 stages instead of 2 (ablation) | SpiralNet++ 3DMM · COMA K=9 | mean error (mm): 0.389 -> 0.373 | Improved |
| add adaptive operator to all 4 stages (ablation) | SpiralNet++ 3DMM · COMA K=9 | mean error (mm): 0.373 -> 0.345 | Improved |
| add static depthwise layer W_k to SpiralNet++ (ablation) | SpiralNet++ · COMA reconstruction loss | reconstruction loss | No clear effect |
| add spatial adaptivity, dynamic weights and global receptive field progressively (ablation) | SpiralNet++ · COMA reconstruction loss | reconstruction loss | Improved |
| naively increase channels instead of adaptive processing | SpiralConv · FAUST correspondence | convergence | Hurt |
| use adaptive spirals instead of static SpiralConv++ on remeshed data | SpiralConv++ · SCAPE remeshed | convergence | Improved |
Closest papers by meaning
- Neural 3D Morphable Models: Spiral Convolutional Networks for 3D Shape Representation Learning and Generation (ICCV 2019 · similarity 0.91)
- Locally Adaptive Neural 3D Morphable Models (CVPR 2024 · similarity 0.87)
- MeshGPT: Generating Triangle Meshes with Decoder-Only Transformers (CVPR 2024 · similarity 0.87)
- Learning Local Neighboring Structure for Robust 3D Shape Representation (AAAI 2021 · similarity 0.86)
- MeshFormer : High-Quality Mesh Generation with 3D-Guided Reconstruction Model (NeurIPS 2024 · similarity 0.85)
- GALA: Geometry-Aware Local Adaptive Grids for Detailed 3D Generation (ICLR 2025 · 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.
