Research map / 3D vision / Face and shape
LIST: Learning Implicitly from Spatial Transformers for Single-View 3D Reconstruction
Mohammad Samiul Arshad, William J. Beksi
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 probabilistic occupancy and occupancy loss (ablation) | LIST Base · ShapeNet 5 classes | CD (x1e-3): 11.35 -> 9.64 | Improved |
| use a single encoder for coarse prediction and query localization (ablation) | LIST · ShapeNet 5 classes | CD (x1e-3): 9.64 -> 10.72 | Hurt |
| use separate encoders and fuse features rather than add two decoder predictions (ablation) | LIST 2D · ShapeNet 5 classes | CD (x1e-3): 8.48 -> 7.32 | Improved |
| use learned localization module instead of estimated camera parameters (ablation) | LIST · ShapeNet 5 classes | CD (x1e-3): 7.89 -> 7.32 | Improved |
| LIST vs D2IM-Net on occluded surfaces | LIST · ShapeNet 5 classes, occluded surfaces | CD_os (x1e-3): 19.43 -> 15.67 | Improved |
| LIST vs IM3D on real Pix3D images | LIST · Pix3D | CD (x1e-3): 6.72 -> 4.36 | Improved |
| use ground-truth camera data instead of estimated for baselines | D2IM-Net style baselines · ShapeNet | reconstruction quality | Improved |
Closest papers by meaning
- Front2Back: Single View 3D Shape Reconstruction via Front to Back Prediction (CVPR 2020 · similarity 0.89)
- D2IM-Net: Learning Detail Disentangled Implicit Fields From Single Images (CVPR 2021 · similarity 0.88)
- Topologically-Aware Deformation Fields for Single-View 3D Reconstruction (CVPR 2022 · similarity 0.88)
- Implicit Functions in Feature Space for 3D Shape Reconstruction and Completion (CVPR 2020 · similarity 0.87)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.87)
- Large Spatial Model: End-to-end Unposed Images to Semantic 3D (NeurIPS 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.
