HumanEdit：一个面向基于指令的图像编辑的高质量人工奖励数据集
HumanEdit: A High-Quality Human-Rewarded Dataset for Instruction-based Image Editing
AI总结：
本文提出 HumanEdit 数据集，通过多阶段人工标注与反馈构建 5,751 张图像及六类编辑指令，支持掩码和无掩码编辑，为指令式图像编辑提供高质量、多样化的新基准。
摘要：
我们提出了 HumanEdit，这是一个高质量、经人工奖励的数据集，专为指令引导的图像编辑而设计，能够通过开放形式的语言指令实现精确且多样化的图像操作。以往的大规模编辑数据集通常只包含极少的人工反馈，导致数据集难以与人类偏好对齐。HumanEdit 通过聘请人工标注者构建数据对，并由管理员提供反馈来弥合这一差距。经过精心整理，HumanEdit 包含 5,751 张图像，并在四个阶段中投入了超过 2,500 小时的人工工作，从而确保其在广泛的图像编辑任务中具有准确性和可靠性。该数据集包含六种不同类型的编辑指令：Action、Add、Counting、Relation、Remove 和 Replace，涵盖了广泛的现实世界场景。数据集中的所有图像都配有掩码，并且对于其中一部分数据，我们确保指令足够详细，以支持无掩码编辑。此外，HumanEdit 提供了全面的多样性以及来自多个领域的高分辨率 $1024 \ imes 1024$ 内容，为指令式图像编辑数据集树立了新的多功能基准。为了推动未来研究并在图像编辑领域建立评估基准，我们在 https://huggingface.co/datasets/BryanW/HumanEdit 发布了 HumanEdit。
英文摘要：
We present HumanEdit, a high-quality, human-rewarded dataset specifically designed for instruction-guided image editing, enabling precise and diverse image manipulations through open-form language instructions. Previous large-scale editing datasets often incorporate minimal human feedback, leading to challenges in aligning datasets with human preferences. HumanEdit bridges this gap by employing human annotators to construct data pairs and administrators to provide feedback. With meticulously curation, HumanEdit comprises 5,751 images and requires more than 2,500 hours of human effort across four stages, ensuring both accuracy and reliability for a wide range of image editing tasks. The dataset includes six distinct types of editing instructions: Action, Add, Counting, Relation, Remove, and Replace, encompassing a broad spectrum of real-world scenarios. All images in the dataset are accompanied by masks, and for a subset of the data, we ensure that the instructions are sufficiently detailed to support mask-free editing. Furthermore, HumanEdit offers comprehensive diversity and high-resolution $1024 \times 1024$ content sourced from various domains, setting a new versatile benchmark for instructional image editing datasets. With the aim of advancing future research and establishing evaluation benchmarks in the field of image editing, we release HumanEdit at https://huggingface.co/datasets/BryanW/HumanEdit.
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。
