CoCoEdit: Content-Consistent Image Editing via Region Regularized Reinforcement Learning

TL;DR

CoCoEdit is a reinforcement learning framework that enhances image editing by ensuring content consistency and precise region control, outperforming existing models in quality and stability.

Contribution

The paper introduces a novel region regularized reinforcement learning approach with pixel-level rewards and dataset augmentation for content-consistent image editing.

Findings

Achieves competitive editing scores with state-of-the-art models.

Significantly improves content consistency as measured by PSNR/SSIM.

Outperforms existing models in human subjective ratings.

Abstract

Image editing has achieved impressive results with the development of large-scale generative models. However, existing models mainly focus on the editing effects of intended objects and regions, often leading to unwanted changes in unintended regions. We present a post-training framework for Content-Consistent Editing (CoCoEdit) via region regularized reinforcement learning. We first augment existing editing datasets with refined instructions and masks, from which 40K diverse and high quality samples are curated as training set. We then introduce a pixel-level similarity reward to complement MLLM-based rewards, enabling models to ensure both editing quality and content consistency during the editing process. To overcome the spatial-agnostic nature of the rewards, we propose a region-based regularizer, aiming to preserve non-edited regions for high-reward samples while encouraging editing effects for low-reward samples. For evaluation, we annotate editing masks for GEdit-Bench and ImgEdit-Bench, introducing pixel-level similarity metrics to measure content consistency and editing quality. Applying CoCoEdit to Qwen-Image-Edit and FLUX-Kontext, we achieve not only competitive editing scores with state-of-the-art models, but also significantly better content consistency, measured by PSNR/SSIM metrics and human subjective ratings.