VGGT-Edit: Feed-forward Native 3D Scene Editing with Residual Field Prediction

2026-05-14

AI Summary

Existing feedforward 3D scene editing methods struggle to respond to dynamic textual instructions, while 2D-lifting strategies often suffer from texture blurriness and geometric inconsistencies. This work proposes the first native 3D feedforward editing framework, which aligns semantic instructions with spatial structure through a depth-synchronized text injection mechanism and employs a residual transformation head to directly predict 3D geometric displacements. To support training, we introduce DeltaScene, a high-quality dataset generated via a pipeline combining multi-objective losses and automatic 3D consistency filtering. Experiments demonstrate that our approach significantly outperforms 2D-lifting baselines in object detail sharpness and multi-view consistency, while achieving near real-time inference speed.

Abstract

High-quality 3D scene reconstruction has recently advanced toward generalizable feed-forward architectures, enabling the generation of complex environments in a single forward pass. However, despite their strong performance in static scene perception, these models remain limited in responding to dynamic human instructions, which restricts their use in interactive applications. Existing editing methods typically rely on a 2D-lifting strategy, where individual views are edited independently and then lifted back into 3D space. This indirect pipeline often leads to blurry textures and inconsistent geometry, as 2D editors lack the spatial awareness required to preserve structure across viewpoints. To address these limitations, we propose VGGT-Edit, a feed-forward framework for text-conditioned native 3D scene editing. VGGT-Edit introduces depth-synchronized text injection to align semantic guidance with the backbone's spatial poses, ensuring stable instruction grounding. This semantic signal is then processed by a residual transformation head, which directly predicts 3D geometric displacements to deform the scene while preserving background stability. To ensure high-fidelity results, we supervise the framework with a multi-term objective function that enforces geometric accuracy and cross-view consistency. We also construct the DeltaScene Dataset, a large-scale dataset generated through an automated pipeline with 3D agreement filtering to ensure ground-truth quality. Experiments show that VGGT-Edit substantially outperforms 2D-lifting baselines, producing sharper object details, stronger multi-view consistency, and near-instant inference speed.

Problem

• 3D scene editing
• feed-forward architecture
• cross-view consistency
• text-conditioned editing
• geometric inconsistency

Innovation

• feed-forward 3D editing
• residual field prediction
• depth-synchronized text injection
• native 3D scene editing
• multi-view consistency