Feedforward 3D Editing Learns from Semantic-Part Transformation

• URL: http://arxiv.org/abs/2605.27351v2
• Date: Wed, 27 May 2026 16:05:26 GMT
• Title: Feedforward 3D Editing Learns from Semantic-Part Transformation
• Abstract summary: A central challenge of feedforward 3D editing lies in the lack of high-quality paired supervision.<n>We introduce a new perspective: scalable feedforward 3D editing should be learned from semantic-part transformations.<n>PartFlow is a feedforward 3D editing network that injects source-aware latent control into pretrained 3D generative priors.
• Score: 13.61340944347895
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: 3D editing is a fundamental capability for scalable 3D content creation. While image editing has rapidly evolved toward large-scale feedforward generative paradigms, 3D AI generation remains dominated by training-free editing pipelines. A central challenge of feedforward 3D editing lies in the lack of high-quality paired supervision. Editable 3D assets require simultaneous preservation of geometry, multi-view consistency, structural coherence, and localized edit controllability. Existing 3D editing datasets often rely on independently generated assets, image-mediated reconstruction or narrow edit taxonomies, leading to inaccurate localization, weak preservation, blurred edit boundaries, and limited semantic consistency. In this work, we introduce a new perspective: scalable feedforward 3D editing should be learned from semantic-part transformations. Based on this insight, we propose Pxform, a high-quality 3D editing dataset with over 100K consistent before/after editing pairs across seven edit types. Instead of treating objects as unstructured shapes, our pipeline grounds edits directly in semantic 3D parts. Built upon Pxform, we further propose PartFlow, a feedforward 3D editing network that injects source-aware latent control into pretrained 3D generative priors. PartFlow introduces mask-aware velocity preservation and render-space consistency supervision to jointly improve edit fidelity and source preservation, while requiring no 3D edit mask during inference. Extensive experiments demonstrate that high-quality semantic-part supervision substantially improves scalable 3D editing, enabling PartFlow to achieve state-of-the-art performance on both geometric and appearance editing benchmarks.

Related papers

• EditVerse3D: High-Quality 3D Object Editing with Region-Aware Learning [59.36059865997641]
  We propose EditVerse3D, a novel 3D editing framework that enables high-quality object editing under coarse guidance.<n>Our approach takes as input a 3D object to be edited, a coarse 3D bounding box indicating the target region, and a reference 2D image describing the desired modification.<n>To facilitate this editing, we introduce a novel region-aware adaptive loss that emphasizes hard-to-learn regions.
  arXiv  Detail & Related papers  (2026-07-08T09:21:22Z)
• Geometry-Instructed Video Editing [54.86319853573801]
  GIVE is a geometry-instructed video editing framework that represents edits through a unified object-state formulation.<n>We build a scalable graphics-engine pipeline that executes object-level edit programs and renders controlled before/after pairs.<n> Experimental results demonstrate that GIVE produces faithful geometric edits with temporal coherence and consistent secondary effects.
  arXiv  Detail & Related papers  (2026-06-23T07:11:45Z)
• JointEdit3D: Feed-Forward 3D Scene Editing in a Unified Latent Space [14.944378716099422]
  Existing 3D scene editing methods rely on per-scene optimization over explicit 3D representations or cascaded edit-and-reconstruct pipelines.<n>We build on a unified RGB-geometry reconstruction-generation latent space and adapt it to feed-forward 3D scene editing.<n>JointEdit3D performs asymmetric latent inpainting by observing only a single edited RGB reference latent.
  arXiv  Detail & Related papers  (2026-06-11T13:35:23Z)
• VGGT-Edit: Feed-forward Native 3D Scene Editing with Residual Field Prediction [59.303842406260124]
  VGGT-Edit is a feed-forward framework for text-conditioned native 3D scene editing.<n>We introduce depth-synchronized text injection to align semantic guidance with the backbone's spatial poses.<n> VGGT-Edit substantially outperforms 2D-lifting baselines, producing sharper object details, stronger multi-view consistency, and near-instant inference speed.
  arXiv  Detail & Related papers  (2026-05-14T17:59:04Z)
• Easy3E: Feed-Forward 3D Asset Editing via Rectified Voxel Flow [29.8200628539749]
  We propose an effective and fully feedforward 3D editing framework based on the TRELLIS generative backbone.<n>Our framework addresses two key issues: adapting training-free 2D editing to structured 3D representations, and overcoming the bottleneck of appearance fidelity in compressed 3D features.
  arXiv  Detail & Related papers  (2026-02-25T02:15:14Z)
• Vinedresser3D: Agentic Text-guided 3D Editing [26.81659566314386]
  Vinedresser3D is an agentic framework for high-quality text-guided 3D editing.<n>It operates directly in the latent space of a native 3D generative model.<n>Experiments on diverse 3D edits demonstrate that Vinedresser3D outperforms prior baselines in both automatic metrics and human preference studies.
  arXiv  Detail & Related papers  (2026-02-23T06:30:36Z)
• NANO3D: A Training-Free Approach for Efficient 3D Editing Without Masks [29.825760228576744]
  Nano3D is a training-free framework for precise and coherent 3D object editing without masks.<n>We construct the first large-scale 3D editing datasets Nano3D-Edit-100k, which contains over 100,000 high-quality 3D editing pairs.
  arXiv  Detail & Related papers  (2025-10-16T17:51:50Z)
• Towards Scalable and Consistent 3D Editing [32.16698854719098]
  3D editing has wide applications in immersive content creation, digital entertainment, and AR/VR.<n>Unlike 2D editing, it remains challenging due to the need for cross-view consistency, structural fidelity, and fine-grained controllability.<n>We introduce 3DEditVerse, the largest paired 3D editing benchmark to date, comprising 116,309 high-quality training pairs and 1,500 curated test pairs.<n>On the model side, we propose 3DEditFormer, a 3D-structure-preserving conditional transformer.
  arXiv  Detail & Related papers  (2025-10-03T13:34:55Z)
• PrEditor3D: Fast and Precise 3D Shape Editing [100.09112677669376]
  We propose a training-free approach to 3D editing that enables the editing of a single shape within a few minutes.<n>The edited 3D mesh aligns well with the prompts, and remains identical for regions that are not intended to be altered.
  arXiv  Detail & Related papers  (2024-12-09T15:44:47Z)
• View-Consistent 3D Editing with Gaussian Splatting [50.6460814430094]
  View-consistent Editing (VcEdit) is a novel framework that seamlessly incorporates 3DGS into image editing processes.<n>By incorporating consistency modules into an iterative pattern, VcEdit proficiently resolves the issue of multi-view inconsistency.
  arXiv  Detail & Related papers  (2024-03-18T15:22:09Z)
• Plasticine3D: 3D Non-Rigid Editing with Text Guidance by Multi-View Embedding Optimization [21.8454418337306]
  We propose Plasticine3D, a novel text-guided controlled 3D editing pipeline that can perform 3D non-rigid editing. Our work divides the editing process into a geometry editing stage and a texture editing stage to achieve separate control of structure and appearance. For the purpose of fine-grained control, we propose Embedding-Fusion (EF) to blend the original characteristics with the editing objectives in the embedding space.
  arXiv  Detail & Related papers  (2023-12-15T09:01:54Z)
• SHAP-EDITOR: Instruction-guided Latent 3D Editing in Seconds [73.91114735118298]
  Shap-Editor is a novel feed-forward 3D editing framework. We demonstrate that direct 3D editing in this space is possible and efficient by building a feed-forward editor network.
  arXiv  Detail & Related papers  (2023-12-14T18:59:06Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.