Room-to-Room (R2R) dataset

doi:doi:10.57702/e4wqnskj

Room-to-Room (R2R) dataset

The Room-to-Room (R2R) dataset is a benchmark for vision-and-language navigation tasks. It consists of 7,189 paths sampled from its navigation graphs, each with three ground-truth navigation instructions written by humans.

BibTex:

@dataset{Chih-Yao_Ma_and_Jiasen_Lu_and_Zuxuan_Wu_and_Ghassan_AlRegib_and_Zsolt_Kira_and_Richard_Socher_and_Caiming_Xiong_2024,
    abstract = {The Room-to-Room (R2R) dataset is a benchmark for vision-and-language navigation tasks. It consists of 7,189 paths sampled from its navigation graphs, each with three ground-truth navigation instructions written by humans.},
    author = {Chih-Yao Ma and Jiasen Lu and Zuxuan Wu and Ghassan AlRegib and Zsolt Kira and Richard Socher and Caiming Xiong},
    doi = {10.57702/e4wqnskj},
    institution = {No Organization},
    keyword = {'3D Indoor Environments', '3D environments', 'Natural Language Instructions', 'Navigation', 'Navigation Instructions', 'Panoramic views', 'Paths', 'Room-to-Room', 'Vision-Language Navigation', 'building-scale 3D environments', 'instruction-route pairs', 'navigation', 'panoramic scenes', 'question answering', 'real-world navigation', 'room-to-room', 'room-to-room navigation', 'single entity', 'single relation', 'vision-and-language'},
    month = {dec},
    publisher = {TIB},
    title = {Room-to-Room (R2R) dataset},
    url = {https://service.tib.eu/ldmservice/dataset/room-to-room--r2r--dataset},
    year = {2024}
}