Summary
Researchers Anirudh Sundara Rajan, Krishna Kumar Singh, and Yong Jae Lee propose a framework for long-horizon image editing in computer vision. Modern image editing models produce realistic results but struggle with abstract, multi-step instructions such as 'make this advertisement more vegetarian-friendly'. Prior agent-based methods decompose such tasks but rely on handcrafted pipelines or teacher imitation, which limits flexibility and decouples learning from actual editing outcomes. The proposed experiential framework couples a planner that generates structured atomic decompositions of instructions with an orchestrator that selects tools and image 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 fed back to refine the planner. The paper is available on arXiv (2605.15181).
Paper Overview
Field: Computer Vision (CV)
Authors: Anirudh Sundara Rajan, Krishna Kumar Singh, Yong Jae Lee
arXiv: 2605.15181
Abstract
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.
The authors propose an experiential framework for long-horizon image editing, where:
- A planner generates structured atomic decompositions of high-level instructions.
- 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 learning, the framework adapts to open-ended editing tasks rather than following fixed pipelines.
Links
- arXiv page: https://arxiv.org/abs/2605.15181
*Auto-collected on 2026-05-15.*
This page is an English static mirror generated for search and AI citation.
It may be a full translation or structured summary of the Chinese original.
Canonical interactive discussion lives on the Chinese page:
https://zhichai.net/topic/177620064