Research map / 3D vision / Face and shape
Deep Implicit Templates for 3D Shape Representation
Check 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 |
|---|---|---|---|
| factor the implicit representation into a shared template and conditional spatial warping | DeepSDF-style shape representation · ShapeNet reconstruction, especially airplanes, sofas, and cars | CD Mean: Airplanes 0.07 -> 0.053; Sofas 0.11 -> 0.093; Cars 0.06 -> 0.052 | Helped |
| factor the implicit representation into a shared template and conditional spatial warping | DeepSDF-style shape representation · ShapeNet Chairs | CD Mean: 0.16 -> 0.20 | Hurt |
| factor the implicit representation into a shared template and conditional spatial warping | 3D shape correspondence · Planes and Cars keypoints | PCK: Planes: 0.411 / 0.628 vs 0.365 / 0.652; Cars: 0.262 / 0.402 vs 0.345 / 0.530 | Mixed |
| add progressive reconstruction loss at multiple warping steps (ablation) | Deep Implicit Templates · ShapeNet Airplanes | CD Mean: 0.042 -> 0.025 | Helped |
| add progressive reconstruction loss at multiple warping steps (ablation) | Deep Implicit Templates · ShapeNet Sofas | CD Mean: 0.117 -> 0.093 | Helped |
| add progressive reconstruction loss at multiple warping steps (ablation) | Deep Implicit Templates · ShapeNet Chairs | CD Mean: 0.23 -> 0.20 | Helped |
| add point-pair regularization (ablation) | Deep Implicit Templates · Planes keypoints (PCK threshold 0.01) | PCK: 0.306 -> 0.365 | Helped |
| add point-pair regularization (ablation) | Deep Implicit Templates · Cars keypoints (PCK threshold 0.01) | PCK: 0.305 -> 0.345 | Helped |
| replace Spatial Warping LSTM with an MLP that predicts the transformation in one step (ablation) | Deep Implicit Templates · latent-space shape interpolation | generalization | Hurt |
| remove point-pair regularization (ablation) | Deep Implicit Templates · ShapeNet reconstruction | reconstruction accuracy | No clear effect |
Closest papers by meaning
- Deformed Implicit Field: Modeling 3D Shapes With Learned Dense Correspondence (CVPR 2021 · similarity 0.91)
- Local Deep Implicit Functions for 3D Shape (CVPR 2020 · similarity 0.90)
- Topology-Preserving Shape Reconstruction and Registration via Neural Diffeomorphic Flow (CVPR 2022 · similarity 0.89)
- LP-DIF: Learning Local Pattern-Specific Deep Implicit Function for 3D Objects and Scenes (CVPR 2023 · similarity 0.89)
- Implicit Modeling of Non-rigid Objects with Cross-Category Signals (AAAI 2024 · similarity 0.88)
- Diffusion 3D Features (Diff3F): Decorating Untextured Shapes with Distilled Semantic Features (CVPR 2024 · similarity 0.87)
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.