TanGO: Training-Free 3D Editing via Tangent-Space Guidance and Optimization
For 3D generative model users, TanGO enables more precise and artifact-free editing without retraining, addressing a key limitation of current flow-matching models.
TanGO introduces a training-free framework for 3D editing that uses tangent-space guidance to adaptively steer per-token modifications, reducing structural artifacts and achieving state-of-the-art performance over existing baselines.
While recent flow-matching 3D generative models (e.g., VecSet) adopt structured representations, their tokens share global context, causing conventional training-free editing to suffer from semantic artifacts such as collapsed preserved regions or incomplete transformations. To address this, we propose TanGO, a training-free framework that enables adaptive per-token steering in the tangent space of generative dynamics. To realize this selective control, we formulate a one-step optimal control rule and determine the strength of each token's control signal using a von Mises-Fisher inspired directional discrepancy derived from the source and target velocity fields. Experiments show that TanGO substantially reduces structural artifacts and achieves state-of-the-art performance, outperforming existing 3D editing baselines. The code is publicly available at https://github.com/siw00-lim/TanGO.