Scene Graph to Knowledge Graph Translation

该数据集包含了2247个来自公寓场景图的原始物体，并转换成了一个包含6827个节点和24116条边的知识图谱，以支持机器人操作任务。数据集还包括了物体及其关联关节的变换、几何信息和惯性属性，从而能够基于场景图执行诸如识别物体和确定合适存储位置等任务。该数据集的规模为：场景图中包含2247个原始物体，知识图谱中包含6827个节点，其任务重点是机器人操作与环境互动。

This dataset contains 2,247 raw objects from apartment scene graphs, and is converted into a knowledge graph with 6,827 nodes and 24,116 edges to support robotic manipulation tasks. The dataset also includes transformation, geometric information, and inertial properties of objects and their associated joints, enabling tasks such as object recognition and determination of appropriate storage locations based on the scene graphs. The scale of this dataset is specified as follows: the scene graphs contain 2,247 raw objects, and the knowledge graph includes 6,827 nodes, with its core task focus on robotic manipulation and environment interaction.