AdaEdit: Adaptive Temporal and Channel Modulation for Flow-Based Image Editing

TL;DR

AdaEdit introduces an adaptive, training-free image editing framework that improves feature injection and perturbation strategies, leading to more seamless and effective text-guided image manipulation in flow-based models.

Contribution

It proposes a progressive injection schedule with continuous decay functions and a channel-selective latent perturbation method, addressing the injection dilemma in flow-based image editing.

Findings

Achieves 8.7% reduction in LPIPS, 2.6% in SSIM, and 2.3% in PSNR on PIE-Bench.

Compatible with multiple ODE solvers and maintains competitive CLIP similarity.

Fully plug-and-play framework for improved image editing quality.

Abstract

Inversion-based image editing in flow matching models has emerged as a powerful paradigm for training-free, text-guided image manipulation. A central challenge in this paradigm is the injection dilemma: injecting source features during denoising preserves the background of the original image but simultaneously suppresses the model's ability to synthesize edited content. Existing methods address this with fixed injection strategies -- binary on/off temporal schedules, uniform spatial mixing ratios, and channel-agnostic latent perturbation -- that ignore the inherently heterogeneous nature of injection demand across both the temporal and channel dimensions. In this paper, we present AdaEdit, a training-free adaptive editing framework that resolves this dilemma through two complementary innovations. First, we propose a Progressive Injection Schedule that replaces hard binary cutoffs with continuous decay functions (sigmoid, cosine, or linear), enabling a smooth transition from source-feature preservation to target-feature generation and eliminating feature discontinuity artifacts. Second, we introduce Channel-Selective Latent Perturbation, which estimates per-channel importance based on the distributional gap between the inverted and random latents and applies differentiated perturbation strengths accordingly -- strongly perturbing edit-relevant channels while preserving structure-encoding channels. Extensive experiments on the PIE-Bench benchmark (700 images, 10 editing types) demonstrate that AdaEdit achieves an 8.7% reduction in LPIPS, a 2.6% improvement in SSIM, and a 2.3% improvement in PSNR over strong baselines, while maintaining competitive CLIP similarity. AdaEdit is fully plug-and-play and compatible with multiple ODE solvers including Euler, RF-Solver, and FireFlow. Code is available at https://github.com/leeguandong/AdaEdit