Research map / 3D vision / Face and shape
Dense 3D Object Reconstruction from a Single Depth View
Bo Yang, Stefano Rosa, Andrew Markham, Niki Trigoni, Hongkai Wen
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 adversarial learning with a conditional discriminator (ablation) | 3D-RecAE · Single-category, same views, chair | IoU: 0.641 -> 0.647 | Improved |
| add adversarial learning with a conditional discriminator (ablation) | 3D-RecAE · Single-category, same views, bench | CE loss: 0.036 -> 0.034 | Improved |
| add adversarial learning with a conditional discriminator (ablation) | 3D-RecAE · Single-category, cross views, couch | IoU: 0.639 -> 0.646 | Improved |
| add adversarial learning with a conditional discriminator (ablation) | 3D-RecAE · Multi-category, same views, bench | CE loss: 0.037 -> 0.030 | Improved |
| add adversarial learning with a conditional discriminator (ablation) | 3D-RecAE · Multi-category, cross views, bench | IoU: 0.530 -> 0.540 | Improved |
| add adversarial learning with a conditional discriminator (ablation) | 3D-RecAE · Real-world Kinect, couch | CE loss: 0.318 -> 0.314 | Improved |
| add adversarial learning with a conditional discriminator (ablation) | 3D-RecAE · Cross-category, car | IoU: 0.557 -> 0.555 | Hurt |
| add adversarial learning with a conditional discriminator (ablation) | 3D-RecAE · Multi-category, bench | mean precision: 0.668 -> 0.680 | Improved |
| add adversarial learning with a conditional discriminator (ablation) | 3D-RecAE · Multi-category, bench | mean recall: 0.808 -> 0.804 | No clear effect |
| train on multiple categories instead of single categories | 3D-RecGAN++ · same views, chair | IoU: 0.647 -> 0.640 | No clear effect |
| train only on bench and test on other categories | 3D-RecGAN++ · Cross-category chair/couch | IoU: 0.647 -> 0.510 | Hurt |
| replace Han et al. LSTM pipeline with skip-connected encoder-decoder | Han et al. · Computation | GPU time: 276.4 -> 38.9 | Improved |
Closest papers by meaning
- Progressive Learning of 3D Reconstruction Network From 2D GAN Data (TPAMI 2023 · similarity 0.88)
- CoReNet: Coherent 3D Scene Reconstruction from a Single RGB Image (ECCV 2020 · similarity 0.88)
- Learning to Reconstruct High-quality 3D Shapes with Cascaded Fully Convolutional Networks (ECCV 2018 · similarity 0.88)
- MVPNet: Multi-View Point Regression Networks for 3D Object Reconstruction from A Single Image (AAAI 2019 · similarity 0.88)
- G3DR: Generative 3D Reconstruction in ImageNet (CVPR 2024 · similarity 0.87)
- Decompositional Neural Scene Reconstruction with Generative Diffusion Prior (CVPR 2025 · 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.
