Research map / 3D vision / Face and shape
Multiresolution Deep Implicit Functions for 3D Shape Representation
Zhang Chen, Yinda Zhang, Kyle Genova, Sean Fanello, Sofien Bouaziz, Christian Häne, Ruofei Du, Cem Keskin, Thomas Funkhouser, Danhang Tang
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 multiresolution hierarchy instead of global-only latent (ablation) | MDIF · ShapeNet chair auto-encoding | Chamfer L2: 0.228 -> 0.009 | Improved |
| use multiresolution hierarchy instead of local-only latent grid (ablation) | MDIF · ShapeNet chair shape completion | Chamfer L2: 5.47 -> 1.34 | Improved |
| add global consistency loss (ablation) | MDIF · ShapeNet chair shape completion | Chamfer L2: 1.43 -> 1.34 | Improved |
| decode residuals instead of SDF directly (ablation) | MDIF · ShapeNet chair shape completion | Chamfer L2: 3.00 -> 1.34 | Improved |
| add latent grid dropout (ablation) | MDIF · ShapeNet chair shape completion | Chamfer L2: 8.38 -> 1.34 | Improved |
| add global connection from level 0 to higher levels (ablation) | MDIF · ShapeNet chair shape completion | Chamfer L2: 19.9 -> 1.34 | Improved |
| add global connection (ablation) | MDIF · ShapeNet chair auto-encoding | Chamfer L2: 0.086 -> 0.009 | Improved |
| add latent grid dropout (ablation) | MDIF · ShapeNet chair auto-encoding | Chamfer L2: 0.026 -> 0.009 | Improved |
| use decoder-only latent optimization instead of encoder-decoder inference | MDIF · 3D-R2N2 test auto-encoding | Chamfer L2: 0.19 -> 0.10 | Improved |
| increase from 5 to 8 levels (ablation) | MDIF · 3D-R2N2 test auto-encoding | Chamfer L2: 0.19 -> 0.12 | Improved |
| use MDIF instead of IF-Net | MDIF · point cloud completion | Chamfer L2: 1.61 -> 0.39 | Improved |
| use MDIF instead of IF-Net | MDIF · voxel super-resolution | Chamfer L2: 1.82 -> 0.96 | Improved |
Closest papers by meaning
- Local Deep Implicit Functions for 3D Shape (CVPR 2020 · similarity 0.92)
- LP-DIF: Learning Local Pattern-Specific Deep Implicit Function for 3D Objects and Scenes (CVPR 2023 · similarity 0.90)
- DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation (CVPR 2019 · similarity 0.90)
- Learning Deep Implicit Functions for 3D Shapes With Dynamic Code Clouds (CVPR 2022 · similarity 0.89)
- Deep Implicit Templates for 3D Shape Representation (CVPR 2021 · similarity 0.88)
- Implicit Modeling of Non-rigid Objects with Cross-Category Signals (AAAI 2024 · similarity 0.88)
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.
