Look Before You Edit: Attention-Guided Camera Placement and Multi-View Alignment for 3D Gaussian Splatting Editing

• URL: http://arxiv.org/abs/2607.19777v1
• Date: Wed, 22 Jul 2026 05:51:10 GMT
• Title: Look Before You Edit: Attention-Guided Camera Placement and Multi-View Alignment for 3D Gaussian Splatting Editing
• Abstract summary: We present LB-Edit, a framework that addresses two coupled problems in 3D scene editing.<n>First, Attention-Guided Editing Camera Placement (ACP) probes the diffusion model's self- and cross-attention at multiple candidate camera distances.<n>Second, Multi-view Attention Alignment (MAA) steers the editor toward the same edit across views along two axes.
• Score: 14.247102544387118
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Text-driven 3D scene editing with 3D Gaussian Splatting (3DGS) typically applies a 2D diffusion editor to views rendered from fixed training cameras, limiting both the spatial coverage of edits and the user's freedom to target specific objects in complex scenes. We present LB-Edit, a framework that addresses two coupled problems: where to place editing cameras for localized edits, and how to make per-view edits agree with one another so that the 3D scene remains consistent after fine-tuning. First, Attention-Guided Editing Camera Placement (ACP) probes the diffusion model's self- and cross-attention at multiple candidate camera distances to find where attention is well-contained in the region of interest, then places a compact, geometrically diverse editing camera set at that attention-optimal distance. Second, Multi-view Attention Alignment (MAA) steers the editor toward the same edit across views along two axes: it aligns appearance by sharing self-attention features via token-level correspondence, and aligns spatial location by lifting cross-attention maps onto the 3D Gaussians as a shared 3D attention field, suppressing both appearance and spatial drift. Experiments on multi-object and single-object scenes show that our method achieves the highest user preference in instruction fidelity, multi-view consistency, and editing locality, using as few as 5 editing views and reducing latency by up to 7x over existing methods.

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)
• View-Consistent 3D Scene Editing via Dual-Path Structural Correspondense and Semantic Continuity [11.663455227298122]
  Cross-view inconsistency remains a major bottleneck in text-driven 3D scene editing.<n>We recast multi-view consistent 3D editing from a distributional perspective.<n>We introduce a dual-path consistency mechanism consisting of projection-guided structural guidance and patch-level semantic propagation.
  arXiv  Detail & Related papers  (2026-04-20T04:47:15Z)
• Mastering Regional 3DGS: Locating, Initializing, and Editing with Diverse 2D Priors [67.22744959435708]
  3D semantic parsing often underperforms compared to its 2D counterpart, making targeted manipulations within 3D spaces more difficult and limiting the fidelity of edits.<n>We address this problem by leveraging 2D diffusion editing to accurately identify modification regions in each view, followed by inverse rendering for 3D localization.<n> Experiments demonstrate that our method achieves state-of-the-art performance while delivering up to a $4times$ speedup.
  arXiv  Detail & Related papers  (2025-07-07T19:15:43Z)
• InterGSEdit: Interactive 3D Gaussian Splatting Editing with 3D Geometry-Consistent Attention Prior [3.6657066043195585]
  InterGSEdit is a novel framework for high-quality 3DGS editing via interactively selecting key views with users' preferences.<n>We project $GAP3D$ to obtain 3D-constrained attention, which are fused with 2D cross-attention via Attention Fusion Network (AFN)
  arXiv  Detail & Related papers  (2025-07-07T13:04:26Z)
• Diffusion-Based Attention Warping for Consistent 3D Scene Editing [55.2480439325792]
  We present a novel method for 3D scene editing using diffusion models.<n>Our approach leverages attention features extracted from a single reference image to define the intended edits.<n>Injecting these warped features into other viewpoints enables coherent propagation of edits.
  arXiv  Detail & Related papers  (2024-12-10T23:57:18Z)
• 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)
• DGE: Direct Gaussian 3D Editing by Consistent Multi-view Editing [72.54566271694654]
  We consider the problem of editing 3D objects and scenes based on open-ended language instructions.<n>A common approach to this problem is to use a 2D image generator or editor to guide the 3D editing process.<n>This process is often inefficient due to the need for iterative updates of costly 3D representations.
  arXiv  Detail & Related papers  (2024-04-29T17:59:30Z)
• 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)

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.