Research map / 3D vision / Face and shape
3D-RCNN: Instance-Level 3D Object Reconstruction via Render-and-Compare
Read the paperCheck 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 |
|---|---|---|---|
| Use 3D-RCNN instead of the strongest listed baseline | Pascal3D+ Bicycle viewpoint estimation · Pascal3D+ Bicycle | AVP4: 67.0 -> 74.3 | Helped |
| Use 3D-RCNN instead of the strongest listed baseline | Pascal3D+ Motorcycle viewpoint estimation · Pascal3D+ Motorcycle | AVP24: 37.5 -> 47.1 | Helped |
| Use 3D-RCNN instead of the strongest listed baseline | Pascal3D+ Car viewpoint estimation · Pascal3D+ Car | AVP4: 58.3 -> 71.8 | Helped |
| Use 3D-RCNN (ResNet50) instead of Deep3DBox | Pascal3D+ Car viewpoint estimation with ground-truth detections · Pascal3D+ Car | Accπ/6: 0.90 -> 0.96 | Helped |
| Use 3D-RCNN (ResNet50) instead of Deep3DBox | Pascal3D+ Car viewpoint estimation with ground-truth detections · Pascal3D+ Car | MedErr: 5.8° -> 3.0° | Helped |
| Use 3D-RCNN instead of DeepMANTA | KITTI car orientation estimation · KITTI test Easy | AAE: 2.85° -> 2.42° | Helped |
| Use 3D-RCNN instead of DeepMANTA | KITTI car orientation estimation · KITTI test Moderate | AAE: 4.98° -> 4.54° | Helped |
| Use 3D-RCNN instead of DeepMANTA | KITTI car orientation estimation · KITTI test Hard | AAE: 6.12° -> 6.00° | Helped |
| Render the predicted 3D scene to produce detection boxes rather than retaining the original boxes | 3D-RCNN with fixed KITTI detections · KITTI train/validation split, hard | AP: 72.7% -> 79.9% | Helped |
Closest papers by meaning
- 3DP3: 3D Scene Perception via Probabilistic Programming (NeurIPS 2021 · similarity 0.89)
- Learning 3D Scene Priors With 2D Supervision (CVPR 2023 · similarity 0.89)
- Learning Unsupervised Hierarchical Part Decomposition of 3D Objects From a Single RGB Image (CVPR 2020 · similarity 0.89)
- Know Your Neighbors: Improving Single-View Reconstruction via Spatial Vision-Language Reasoning (CVPR 2024 · similarity 0.88)
- ImageNet3D: Towards General-Purpose Object-Level 3D Understanding (NeurIPS 2024 · similarity 0.88)
- Learning Partonomic 3D Reconstruction from Image Collections (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.