Tabletop Tidying Up (TTU) dataset
收藏资源简介:
TTU数据集是一个结构化的数据集,收集于模拟环境中,包含了整理前后场景的RGB-D图像、对象类别和6自由度位置信息。该数据集由首尔国立大学的研究团队创建,包含咖啡桌、餐桌、办公桌和浴室四种环境,共有170个对象模型,413个模板,以及224,225个场景数据。数据集旨在训练一个视觉基础的整洁度判别器,用于评估桌面物体的整洁度,并用于引导蒙特卡洛树搜索算法进行桌面整理。
The TTU Dataset is a structured dataset collected in a simulated environment. It contains RGB-D images of scenes before and after tidying, object categories, and 6-degree-of-freedom (6DoF) pose information. Developed by a research team from Seoul National University, this dataset covers four environmental settings: coffee tables, dining tables, office desks, and bathrooms, and includes 170 object models, 413 templates, as well as 224,225 scene data samples. The dataset is intended to train a vision-based tidiness discriminator for evaluating the tidiness of tabletop objects, and to guide Monte Carlo Tree Search (MCTS) algorithms for tabletop tidying tasks.




