Research map / 3D vision / Face and shape
Single-View 3D Object Reconstruction From Shape Priors in Memory
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 |
|---|---|---|---|
| Retrieve related shape priors from memory and combine them with image features | Mem3D · ShapeNet | IoU: 0.670 -> 0.729 | Helped |
| Retrieve related shape priors from memory and combine them with image features | Mem3D · ShapeNet | F-Score@1%: 0.436 -> 0.517 | Helped |
| Use memory-based shape priors | Mem3D · Pix3D chair, untruncated and unoccluded (2,894 images) | IoU: 0.325 -> 0.387 | Helped |
| Use memory-based shape priors | Mem3D · Pix3D chair, untruncated and unoccluded (2,894 images) | F-Score@1%: 0.068 -> 0.143 | Helped |
| Use memory-based shape priors | Mem3D · Pix3D chair, heavy occlusions (945 images) | IoU: 0.215 -> 0.336 | Helped |
| Use memory-based shape priors | Mem3D · Pix3D chair, heavy occlusions (945 images) | F-Score@1%: 0.041 -> 0.105 | Helped |
| Add a memory network containing image-voxel shape priors (ablation) | Mem3D · Pix3D chair, untruncated and unoccluded (2,894 images) | IoU: 0.273 -> 0.387 | Helped |
| Use memory size m = 1000 instead of m = 4000 (ablation) | Mem3D · Pix3D chair, untruncated and unoccluded (2,894 images) | IoU: 0.366 vs 0.387 | Hurt |
| Use memory size m = 2000 instead of m = 4000 (ablation) | Mem3D · Pix3D chair, untruncated and unoccluded (2,894 images) | IoU: 0.372 vs 0.387 | Hurt |
| Add voxel triplet loss to train image features for shape-based retrieval (ablation) | Mem3D · Pix3D chair, untruncated and unoccluded (2,894 images) | IoU: 0.359 -> 0.387 | Helped |
| Use the top-1 retrieved shape directly rather than encoding retrieved shapes (ablation) | Mem3D · Pix3D chair, untruncated and unoccluded (2,894 images) | IoU: 0.287 vs 0.387 | Hurt |
| Average-fuse retrieved shapes instead of using the LSTM shape encoder (ablation) | Mem3D · Pix3D chair, untruncated and unoccluded (2,894 images) | IoU: 0.363 vs 0.387 | Hurt |
Closest papers by meaning
- Front2Back: Single View 3D Shape Reconstruction via Front to Back Prediction (CVPR 2020 · similarity 0.89)
- Learning to Reconstruct Shapes from Unseen Classes (NeurIPS 2018 · similarity 0.89)
- Pixels, Voxels, and Views: A Study of Shape Representations for Single View 3D Object Shape Prediction (CVPR 2018 · similarity 0.89)
- Learning View Priors for Single-View 3D Reconstruction (CVPR 2019 · similarity 0.89)
- Learning Partonomic 3D Reconstruction from Image Collections (CVPR 2025 · similarity 0.88)
- MESC-3D:Mining Effective Semantic Cues for 3D Reconstruction from a Single Image (CVPR 2025 · 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.