From Plans to Pixels: Learning to Plan and Orchestrate for Open-Ended Image Editing
For users needing complex image editing, this work addresses the limitation of existing models in handling abstract, multi-step instructions by learning from editing outcomes rather than relying on handcrafted pipelines.
The paper tackles abstract, multi-step image editing by introducing an experiential framework where a planner decomposes instructions into atomic steps and an orchestrator selects tools and regions, trained via outcome-based rewards. The approach yields more coherent edits than single-step or rule-based multistep baselines.
Modern image editing models produce realistic results but struggle with abstract, multi step instructions (e.g., ``make this advertisement more vegetarian-friendly''). Prior agent based methods decompose such tasks but rely on handcrafted pipelines or teacher imitation, limiting flexibility and decoupling learning from actual editing outcomes. We propose an experiential framework for long-horizon image editing, where a planner generates structured atomic decompositions and an orchestrator selects tools and regions to execute each step. A vision language judge provides outcome-based rewards for instruction adherence and visual quality. The orchestrator is trained to maximize these rewards, and successful trajectories are used to refine the planner. By tightly coupling planning with reward driven execution, our approach yields more coherent and reliable edits than single-step or rule-based multistep baselines.