Research map / 3D vision / Face and shape
Semi-Supervised Single-View 3D Reconstruction via Prototype Shape Priors
Zhen Xing, Hengduo Li, Zuxuan Wu, Yu-Gang Jiang
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 |
|---|---|---|---|
| add prototype attentive module (ablation) | SSP3D · ShapeNet 32^3 (1% labels) | mean IoU: 43.44 -> 46.99 | Improved |
| add prototype attentive module (ablation) | SSP3D · Pix3D 128^3 (10% labels) | mean IoU: 31.32 -> 35.39 | Improved |
| use attention instead of average fusion of prototypes (ablation) | SSP3D · ShapeNet 32^3 (1% labels) | mean IoU: 43.80 -> 46.99 | Improved |
| use attention instead of LSTM fusion (ablation) | SSP3D · ShapeNet 32^3 (1% labels) | mean IoU: 44.61 -> 46.99 | Improved |
| add shape naturalness module (ablation) | SSP3D · ShapeNet 32^3 (1% labels) | mean IoU: 45.87 -> 46.99 | Improved |
| add discriminator confidence score weighting (ablation) | SSP3D · ShapeNet 32^3 (1% labels) | mean IoU: 46.32 -> 46.99 | Improved |
| add discriminator confidence score weighting (ablation) | SSP3D · Pix3D 128^3 (10% labels) | mean IoU: 34.62 -> 35.39 | Improved |
| use L2 unsupervised loss instead of BCE (ablation) | SSP3D · ShapeNet 32^3 (1% labels) | mean IoU: 45.85 -> 46.99 | Improved |
| add SSP3D semi-supervised training | Pix2Vox · ShapeNet 32^3 (1% labels) | mean IoU: 41.13 -> 46.99 | Improved |
| extend MixMatch to 3D reconstruction | Pix2Vox · ShapeNet 32^3 (10% labels) | mean IoU: 53.99 -> 52.62 | Hurt |
| extend FixMatch to Pix3D | Pix2Vox · Pix3D 128^3 (10% labels) | mean IoU: 29.80 -> 30.35 | Mixed |
| add SSP3D with 10 labeled novel-category shapes | SSP3D · ShapeNet novel categories (10-shot) | mean IoU: 36 -> 48 | Improved |
Closest papers by meaning
- Neural View Synthesis and Matching for Semi-Supervised Few-Shot Learning of 3D Pose (NeurIPS 2021 · similarity 0.89)
- Self-supervised Single-view 3D Reconstruction via Semantic Consistency (ECCV 2020 · similarity 0.89)
- DQS3D: Densely-matched Quantization-aware Semi-supervised 3D Detection (ICCV 2023 · similarity 0.89)
- Few-Shot Generalization for Single-Image 3D Reconstruction via Priors (ICCV 2019 · similarity 0.89)
- Semi-Supervised Monocular 3D Face Reconstruction With End-to-End Shape-Preserved Domain Transfer (ICCV 2019 · similarity 0.89)
- Weakly-Supervised Learning of Category-Specific 3D Object Shapes (TPAMI 2019 · 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.
