VisionKG

VisionKG是由柏林技术大学开放分布式系统团队创建的一个视觉数据集知识图谱，旨在通过知识图谱技术统一管理和访问多样化的视觉数据集。该数据集包含约5.19亿RDF三元组，描述了约4000万实体，涵盖30个数据集和四个流行的计算机视觉任务。VisionKG通过标准化SPARQL查询提供数据访问，支持用户在不同场景下快速访问和查询所需图像数据，有效解决数据不一致和偏差问题，适用于构建和测试视觉识别系统。

VisionKG is a visual dataset knowledge graph developed by the Open Distributed Systems Team at Technische Universität Berlin, aiming to uniformly manage and access diverse visual datasets through knowledge graph technologies. It contains approximately 519 million RDF triples, describing about 40 million entities, and covers 30 datasets along with four popular computer vision tasks. VisionKG provides data access via standardized SPARQL queries, allowing users to quickly access and query required image data across different scenarios. It effectively resolves issues of data inconsistency and bias, and is applicable to the construction and testing of visual recognition systems.