Research map / Image generation / Editing and style transfer
TanGO: Training-Free 3D Editing via Tangent-Space Guidance and Optimization
Siwoo Lim, Sunjae Yoon, Gwanhyeong Koo, Hyeonseo Yun, Chang D. Yoo
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 |
|---|---|---|---|
| replace global scaling with per-token directional steering | AnchorFlow baseline · TanGOEdit | CLIP-T: 0.2124 -> 0.2301 | Improved |
| replace global scaling with per-token directional steering | AnchorFlow baseline · TanGOEdit | DINO-I: 0.6426 -> 0.6510 | Improved |
| replace global scaling with per-token directional steering | AnchorFlow baseline · TanGOEdit | CLIP-I: 0.7492 -> 0.7679 | Improved |
| add mean-gain normalization (vs fixed constant lambda) (ablation) | TanGO · qualitative examples | edit completeness | Improved |
| use directional discrepancy gain instead of global, L2-magnitude or inner-product gains (ablation) | TanGO · TanGOEdit | editing quality | Improved |
| apply TanGO to TripoSG instead of Hunyuan3D 2.1 (ablation) | TanGO · TripoSG | CLIP/3D preservation metrics | Improved |
| combine TanGO with AnchorFlow naively (ablation) | TanGO · qualitative | editing quality | No clear effect |
| increase lambda beyond default (ablation) | TanGO · TanGOEdit | edit strength vs preservation | Mixed |
| use TanGO instead of AnchorFlow (user study) | AnchorFlow · User study | shape preservation preference: 13% -> 87% | Improved |
| use TanGO per-token steering | AnchorFlow · TanGOEdit | inference time: 28.76 -> 28.73 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- AnchorFlow — AnchorFlow: Training-Free 3D Editing via Latent Anchor-Aligned Flows (CVPR 2026) · better on 7 of 7 shared results
- FlowEdit — FlowEdit: Inversion-Free Text-Based Editing Using Pre-Trained Flow Models (ICCV 2025) · better on 3 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Omni-3DEdit: Generalized Versatile 3D Editing in One-Pass (CVPR 2026 · similarity 0.87)
- AnchorFlow: Training-Free 3D Editing via Latent Anchor-Aligned Flows (CVPR 2026 · similarity 0.87)
- Training-Free 3D Editing via Feature-Divergence Localization and Trajectory Correction (NeurIPS 2026 · similarity 0.87)
- Velocity-Space 3D Asset Editing (NeurIPS 2026 · similarity 0.86)
- DreamCatalyst: Fast and High-Quality 3D Editing via Controlling Editability and Identity Preservation (ICLR 2025 · similarity 0.85)
- CTRL-D: Controllable Dynamic 3D Scene Editing with Personalized 2D Diffusion (CVPR 2025 · similarity 0.85)
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.
