WIKIPerson

WIKIPerson是一个高质量的人工标注视觉人物链接数据集，由中国科学院计算技术研究所网络数据科学与技术重点实验室创建。该数据集包含48,000张新闻图片，涵盖13,000个不同的人物实体，每张图片均标注有实体的维基数据ID。数据集通过精心筛选的新闻图片和文本描述，确保了数据的多样性和准确性。WIKIPerson旨在推动视觉命名实体链接技术的发展，特别是在图像检索、图像描述和视觉问答等领域，通过提供详尽的视觉和文本信息，支持多种视觉实体链接任务的研究。

WIKIPerson is a high-quality manually annotated visual person linking dataset, created by the Key Laboratory of Network Data Science and Technology, Institute of Computing Technology, Chinese Academy of Sciences. This dataset contains 48,000 news images covering 13,000 distinct person entities, and each image is annotated with the entity's Wikidata ID. The dataset ensures data diversity and accuracy via carefully screened news images and accompanying textual descriptions. WIKIPerson aims to advance the development of visual named entity linking technologies, particularly in fields such as image retrieval, image captioning, and visual question answering, by providing comprehensive visual and textual information to support research on a variety of visual entity linking tasks.