Model the Edit, Not the Image: Visual Autoregressive Editing from a Source-Centric Perspective
2026-08-10 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors present EditMod, a new way to edit images using a type of AI model called visual autoregressive models (VARs). Unlike previous methods that generate new images from scratch based on text prompts, EditMod focuses on making changes directly to the existing image's internal representation. It compares predictions conditioned on the original and edited text to find the changes needed and applies these to the image data efficiently. Their approach keeps the edited image close to the original while aligning well with the text instructions and works quickly without extra preparation steps.
Visual autoregressive modelsImage editingText-guided editingCoarse-to-fine predictionAutoregressive contextSource-image fidelityScale-wise editingResidual updateToken representationGPU acceleration
Authors
Hongyi Fang, Chuwen Xie, Benjia Zhou, Yu-Xuan Qiu, Chenggong Hu, Zhibin Wang, Chao Chen, Jianbin Qin, Rui Mao
Abstract
Next-scale visual autoregressive models (VARs) have emerged as a powerful generative paradigm, producing high-quality images through efficient coarse-to-fine prediction. However, their potential for text-guided image editing remains largely underexplored. Existing training-free VAR editing approaches often formulate editing as target-conditioned regeneration guided or constrained by the source image, and may rely on inversion, test-time optimization, attention control, or user-provided masks. This generation-centric formulation does not fully exploit the multiscale source representations provided by VARs and may introduce additional computation or intervention. We instead take a source-centric perspective on VAR editing, in which the encoded source image tokens serve as the primary visual state and the editing process focuses on condition-induced changes. Based on this perspective, we propose \textbf{EditMod}, which compares source- and target-conditioned predictions under a shared autoregressive context, treats their difference as a scale-wise editing direction, and applies it as a residual update to source tokens at selected scales. Experiments show that EditMod achieves leading source-image fidelity while maintaining strong text alignment, and completes end-to-end editing of a 1K image in only 1.57 seconds on a single A100 GPU without per-image preparation.