Research map / 3D vision / Face and shape
Fully Convolutional Mesh Autoencoder using Efficient Spatially Varying Kernels
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 |
|---|---|---|---|
| Increase the weight-basis size M from 17 to 37 (ablation) | vcConv/TransConv autoencoder · D-FAUST | test error (mm): 5.01 -> 4.56 | Helped |
| Increase the weight-basis size M from 17 to 27 (ablation) | vcConv/TransConv autoencoder · D-FAUST | test error (mm): 5.01 -> 4.73 | Helped |
| Remove residual layers (ablation) | vcConv/TransConv autoencoder · D-FAUST | test error (mm): 5.01 -> 5.23 | Hurt |
| Replace strided convolution with stride-1 convolution plus vdPool/vdUnpool (ablation) | vcConv/TransConv autoencoder · D-FAUST | test error (mm): 5.01 -> 5.63 | Hurt |
| Normalize the weight bases (ablation) | vcConv/TransConv autoencoder · D-FAUST | test error (mm): 5.01 -> 14.29 | Hurt |
| Replace vcConv with LCConv (ablation) | D-FAUST convolution comparison network · D-FAUST | test error (mm): 5.63 -> 4.23 | Helped |
| Replace vcConv with ChebConv (ablation) | D-FAUST convolution comparison network · D-FAUST | test error (mm): 5.63 -> 8.59 | Hurt |
| Replace vcConv with MoNetConv (ablation) | D-FAUST convolution comparison network · D-FAUST | test error (mm): 5.63 -> 10.4 | Hurt |
| Replace vcConv with GATConv (ablation) | D-FAUST convolution comparison network · D-FAUST | test error (mm): 5.63 -> 14.28 | Hurt |
| Replace vcConv with FeastConv (ablation) | D-FAUST convolution comparison network · D-FAUST | test error (mm): 5.63 -> 17.03 | Hurt |
| Replace vdPool/vdUnpool with average pooling/unpooling (ablation) | vcConv network · D-FAUST | test error (mm): 5.63 -> 6.88 | Hurt |
| Replace vdPool/vdUnpool with max pooling/unpooling (ablation) | vcConv network · D-FAUST | test error (mm): 5.63 -> 13.55 | Hurt |
Closest papers by meaning
- EdgeRunner: Auto-regressive Auto-encoder for Artistic Mesh Generation (ICLR 2025 · similarity 0.87)
- LaGeM: A Large Geometry Model for 3D Representation Learning and Diffusion (ICLR 2025 · similarity 0.87)
- MeshGPT: Generating Triangle Meshes with Decoder-Only Transformers (CVPR 2024 · similarity 0.86)
- Locally Adaptive Neural 3D Morphable Models (CVPR 2024 · similarity 0.86)
- 4Deform: Neural Surface Deformation for Robust Shape Interpolation (CVPR 2025 · similarity 0.86)
- ARMesh: Autoregressive Mesh Generation via Next-Level-of-Detail Prediction (NeurIPS 2025 · 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.