Research map / 3D vision / Face and shape
Shape Unicode: A Unified 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 |
|---|---|---|---|
| Add the cross-representation embedding loss (ablation) | Shape Unicode · ShapeNet test classification | accuracy: 81.61 -> 83.38; 81.77 -> 84.23; 81.11 -> 82.48 | Helped |
| Add cross-representation translation losses (ablation) | Shape Unicode · ShapeNet test classification | accuracy: 81.72 -> 83.38; 82.36 -> 84.23; 79.01 -> 82.48 | Helped |
| Add the joint classification loss (ablation) | Shape Unicode · ShapeNet test classification | accuracy: 81.91 -> 83.38; 82.28 -> 84.23; 81.67 -> 82.48 | Helped |
| Train encoders and decoders jointly across representations rather than independently (ablation) | Shape Unicode · ShapeNet test classification | accuracy: 83.53 -> 83.38; 84.07 -> 84.23; 80.76 -> 82.48 | Mixed |
| Use joint cross-representation codes instead of solo codes (ablation) | Shape Unicode · ShapeNet segmentation | overall segmentation accuracy: 85.73 -> 86.73; 85.05 -> 87.26 | Helped |
| Train a shared code using cross-representation information, then infer from a single input representation (ablation) | Shape Unicode · SHREC'17 retrieval | Micro F1: 0.73 vs 0.68; 0.72 vs 0.72; 0.71 vs 0.68 | Mixed |
| Combine representation predictions with weighted fusion | ShapeNet classification · ShapeNet test | classification accuracy: 81.54 vs 83.38; 81.54 vs 84.23; 81.54 vs 82.48 | Hurt |
| Concatenate codes from all representations | ShapeNet classification · ShapeNet test | classification accuracy: 81.53 vs 83.38; 81.53 vs 84.23; 81.53 vs 82.48 | Hurt |
Closest papers by meaning
- Pixels, Voxels, and Views: A Study of Shape Representations for Single View 3D Object Shape Prediction (CVPR 2018 · similarity 0.89)
- Unsupervised Learning of Intrinsic Structural Representation Points (CVPR 2020 · similarity 0.89)
- $\texttt{NAISR}$: A 3D Neural Additive Model for Interpretable Shape Representation (ICLR 2024 · similarity 0.88)
- Structured Prediction of Unobserved Voxels From a Single Depth Image (CVPR 2016 · similarity 0.88)
- Uni3D: Exploring Unified 3D Representation at Scale (ICLR 2024 · similarity 0.87)
- Mosaic-SDF for 3D Generative Models (CVPR 2024 · 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.