Region-Constrained Group Relative Policy Optimization for Flow-Based Image Editing

TL;DR

This paper introduces RC-GRPO-Editing, a region-constrained framework for flow-based image editing that improves instruction adherence and content preservation by reducing background noise and aligning attention.

Contribution

It proposes a novel region-constrained post-training method that localizes exploration and aligns attention to enhance editing precision in flow-based models.

Findings

Improves editing region instruction adherence.

Enhances non-target content preservation.

Reduces reward variance and stabilizes advantages.

Abstract

Instruction-guided image editing requires balancing target modification with non-target preservation. Recently, flow-based models have emerged as a strong and increasingly adopted backbone for instruction-guided image editing, thanks to their high fidelity and efficient deterministic ODE sampling. Building on this foundation, GRPO-based reward-driven post-training has been explored to directly optimize editing-specific rewards, improving instruction following and editing consistency. However, existing methods often suffer from noisy credit assignment: global exploration also perturbs non-target regions, inflating within-group reward variance and yielding noisy GRPO advantages. To address this, we propose RC-GRPO-Editing, a region-constrained GRPO post-training framework for flow-based image editing under deterministic ODE sampling. It suppresses background-induced nuisance variance to enable cleaner localized credit assignment, improving editing region instruction adherence while preserving non-target content. Concretely, we localize exploration via region-decoupled initial noise perturbations to reduce background-induced reward variance and stabilize GRPO advantages, and introduce an attention concentration reward that aligns cross-attention with the intended editing region throughout the rollout, reducing unintended changes in non-target regions. Experiments on CompBench show consistent improvements in editing region instruction adherence and non-target preservation.