Fukuoka datasets for place categorization
Periodicals(2019)
摘要
AbstractThis paper presents several multi-modal 3D datasets for the problem of categorization of places. In this problem. a robotic agent should decide on the type of place/environment where it is located (residential area, forest, etc.) using information gathered by its sensors. In addition to the 3D depth information, the datasets include additional modalities such as RGB or reflectance images. The observations were taken in different indoor and outdoor environments in Fukuoka city, Japan. Outdoor place categories include forests, urban areas, indoor parking, outdoor parking, coastal areas, and residential areas. Indoor place categories include corridors, offices, study rooms, kitchens, laboratories, and toilets. The datasets are available to download at http://robotics.ait.kyushu-u.ac.jp/kyushu_datasets.
更多查看译文
关键词
place categorization, outdoor scenario, indoor scenario, 3D data, multi-modal data, mobile robotics, labeling, dataset
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络