arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.14916cs.CVcs.MM
收藏

EditHF-1M: 一个百万级的丰富人类偏好反馈用于图像编辑

EditHF-1M: A Million-Scale Rich Human Preference Feedback for Image Editing

Shanghai Jiao Tong University(上海交通大学);Vivo Mobile Communication Co., Ltd(vivo移动通信有限公司);University of Electronic and Science Technology of China(电子科技大学);Ocean University of China(海洋大学);Beijing University of Technology(北京理工大学)

Zitong Xu, Huiyu Duan, Zhongpeng Ji, Xinyun Zhang, Yutao Liu, Xiongkuo Min, Ke Gu, Jian Zhang, Shusong Xu, Jinwei Chen, Bo Li, Guangtao Zhai

AI总结:

本文提出EditHF-1M百万级图像编辑数据集及基于其的EditHF评估模型和EditHF-Reward奖励模型,通过强化学习优化文本引导图像编辑模型,实验表明其在对齐人类偏好和泛化能力方面表现优异。

摘要:

近期文本引导图像编辑(TIE)模型取得了显著进展,但许多编辑图像仍存在伪影、意外编辑和不美观内容等问题。尽管已提出一些评估编辑图像的基准和方法,但可扩展的评估模型仍然缺乏,限制了图像编辑的人类反馈奖励模型的发展。为解决这些挑战,我们首先引入EditHF-1M,一个包含超过2900万对人类偏好和148000个人类平均意见评分的百万级图像编辑数据集,两者均从三个维度评估:视觉质量、指令对齐性和属性保留。基于EditHF-1M,我们提出EditHF,一个基于多模态大语言模型(MLLM)的评估模型,以提供图像编辑的人类对齐反馈。最后,我们引入EditHF-Reward,利用EditHF作为奖励信号,通过强化学习优化文本引导图像编辑模型。大量实验表明,EditHF在对齐人类偏好和在其他数据集上的泛化能力方面表现优异。此外,我们使用EditHF-Reward微调Qwen-Image-Edit,取得了显著的性能提升,这证明了EditHF作为奖励模型的能力,能够扩展图像编辑。数据集和代码将在我们的GitHub仓库中发布:https://github.com/IntMeGroup/EditHF。

英文摘要:

Recent text-guided image editing (TIE) models have achieved remarkable progress, while many edited images still suffer from issues such as artifacts, unexpected editings, unaesthetic contents. Although some benchmarks and methods have been proposed for evaluating edited images, scalable evaluation models are still lacking, which limits the development of human feedback reward models for image editing. To address the challenges, we first introduce \textbf{EditHF-1M}, a million-scale image editing dataset with over 29M human preference pairs and 148K human mean opinion ratings, both evaluated from three dimensions, \textit{i.e.}, visual quality, instruction alignment, and attribute preservation. Based on EditHF-1M, we propose \textbf{EditHF}, a multimodal large language model (MLLM) based evaluation model, to provide human-aligned feedback from image editing. Finally, we introduce \textbf{EditHF-Reward}, which utilizes EditHF as the reward signal to optimize the text-guided image editing models through reinforcement learning. Extensive experiments show that EditHF achieves superior alignment with human preferences and demonstrates strong generalization on other datasets. Furthermore, we fine-tune the Qwen-Image-Edit using EditHF-Reward, achieving significant performance improvements, which demonstrates the ability of EditHF to serve as a reward model to scale-up the image editing. Both the dataset and code will be released in our GitHub repository: https://github.com/IntMeGroup/EditHF.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑