HumanEdit：一個高品質的人類獲得獎勵的基於指示的圖像編輯數據集

HumanEdit: A High-Quality Human-Rewarded Dataset for Instruction-based Image Editing

December 5, 2024

作者: Jinbin Bai, Wei Chow, Ling Yang, Xiangtai Li, Juncheng Li, Hanwang Zhang, Shuicheng Yan

cs.AI

摘要

我們介紹了HumanEdit，這是一個高質量、經人類獎勵的數據集，專門為指導引導的圖像編輯而設計，通過開放形式的語言指令實現精確和多樣化的圖像操作。先前的大規模編輯數據集通常包含最少的人類反饋，導致數據集與人類偏好之間存在挑戰。HumanEdit通過聘請人類標註者構建數據對和管理員提供反饋來彌合這一差距。通過精心策劃，HumanEdit包括5,751張圖像，跨越四個階段需要超過2,500小時的人力，確保廣泛的圖像編輯任務的準確性和可靠性。該數據集包括六種不同類型的編輯指令：動作、添加、計數、關係、移除和替換，涵蓋了廣泛的現實場景。數據集中的所有圖像都附帶蒙版，對於部分數據，我們確保指令足夠詳細以支持無蒙版編輯。此外，HumanEdit提供了全面的多樣性和高分辨率1024x1024內容，來自各個領域，為指導式圖像編輯數據集設定了一個新的多功能基準。為了推進未來研究並在圖像編輯領域建立評估基準，我們在https://huggingface.co/datasets/BryanW/HumanEdit 上發布了HumanEdit。

English

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.