Research map / 3D vision / 3D generation
Real2Edit2Real: Generating Robotic Demonstrations via a 3D Control Interface
Yujie Zhao, Hongwei Fan, Di Chen, Shengcong Chen, Liliang Chen, Xiaoqi Li, Guanghui Ren, Hao Dong
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 |
|---|---|---|---|
| train on 200 generated demos from 1 source demo instead of 10 real demos | Go-1 · Four real-robot tasks | success rate: 36.3% -> 65.0% | Improved |
| increase source demos from 1 to 5 for generation (200 generated) | Go-1 · Four real-robot tasks | success rate: 65.0% -> 78.8% | Improved |
| increase source demos from 1 to 5 for generation (200 generated) | pi0.5 · Four real-robot tasks | success rate: 57.5% -> 81.3% | Improved |
| generated data from 1 source demo vs 50 real demos | pi0.5 · Four real-robot tasks | success rate: 61.3% -> 57.5% | Hurt |
| add generated platform-height demos to training | Go-1 · Height generalization | success rate: 50% -> 80% | Improved |
| generate data with varied desktop textures | Go-1 · Texture variation | success rate: 52% -> 68% | Improved |
| generate data without texture variation vs 50 real demos | Go-1 · Texture variation | success rate: 50% -> 52% | No clear effect |
| add robot pose correction to depth editing (ablation) | Real2Edit2Real · Qualitative (Figure 6) | generation quality | Improved |
| add smooth object relocation (ablation) | Real2Edit2Real · Qualitative (Figure 7) | object placement accuracy | Improved |
| fine-tune VGGT with hybrid real+sim data (Metric-VGGT) (ablation) | VGGT · Qualitative (Figure 5) | point cloud and camera pose quality | Improved |
Closest papers by meaning
- Geometry-aware 4D Video Generation for Robot Manipulation (ICLR 2026 · similarity 0.90)
- ManipDreamer3D: Synthesizing Plausible Robotic Manipulation Video with Occupancy-aware 3D Trajectory (AAAI 2026 · similarity 0.90)
- RoboPearls: Editable Video Simulation for Robot Manipulation (ICCV 2025 · similarity 0.90)
- DynaRend: Learning 3D Dynamics via Masked Future Rendering for Robotic Manipulation (NeurIPS 2025 · similarity 0.89)
- Robots Pre-train Robots: Manipulation-Centric Robotic Representation from Large-Scale Robot Datasets (ICLR 2025 · similarity 0.87)
- GravMAD: Grounded Spatial Value Maps Guided Action Diffusion for Generalized 3D Manipulation (ICLR 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.
