Online Reinforcement Learning (RL) offers a promising avenue for complex image editing but is currently constrained by the scarcity of reliable and fine-grained reward signals. Existing evaluators frequently struggle with a critical perception gap we term "Attention Collapse," where models neglect cross-image comparisons and fail to capture fine-grained details, resulting in inaccurate perception and miscalibrated scores. To address these limitations, we propose SpatialReward, a reward model that enforces precise verification via explicit spatial reasoning. By anchoring reasoning to predicted edit regions, SpatialReward grounds semantic judgments in pixel-level evidence, significantly enhancing evaluative accuracy. Trained on a curated 260k spatial-aware dataset, our model achieves state-of-the-art performance on MMRB2 and EditReward-Bench, and outperforms proprietary evaluators on our proposed MultiEditReward-Bench. Furthermore, SpatialReward serves as a robust signal in online RL, boosting OmniGen2 by +0.90 on GEdit-Bench--surpassing the leading discriminative model and doubling the gain of GPT-4.1 (+0.45). These results demonstrate that spatial reasoning is essential for unlocking effective alignment in image editing.
@article{arxiv.2602.07458,
title = {SpatialReward: Bridging the Perception Gap in Online RL for Image Editing via Explicit Spatial Reasoning},
author = {Yancheng Long and Yankai Yang and Hongyang Wei and Wei Chen and Tianke Zhang and Haonan fan and Changyi Liu and Kaiyu Jiang and Jiankang Chen and Kaiyu Tang and Bin Wen and Fan Yang and Tingting Gao and Han Li and Shuo Yang},
journal= {arXiv preprint arXiv:2602.07458},
year = {2026}
}
Comments
Accepted at the 43rd International Conference on Machine Learning (ICML 2026)