1078 15 5061 4 27306 document 5061 README-1078.txt text/plain 6c3e6761fd3b95d8ac4eeefac6e0c20f MD5 3361 2024-04-08 12:45:41 http://edata.bham.ac.uk/1078/3/README-1078.txt 1078 3 1 text/plain text public cc_by_4

README-1078.txt

readme 5062 1 27310 document 5062 indexcodes.txt text/x-c 94009de21f94310aee98a066e8e62706 MD5 1615 2024-04-08 12:55:52 http://edata.bham.ac.uk/1078/4/indexcodes.txt 1078 4 4 text/x-c other Generate index codes conversion from text to indexcodes public

indexcodes.txt

data http://eprints.org/relation/isVersionOf http://edata.bham.ac.uk/id/document/5061 http://eprints.org/relation/isVolatileVersionOf http://edata.bham.ac.uk/id/document/5061 http://eprints.org/relation/isIndexCodesVersionOf http://edata.bham.ac.uk/id/document/5061 archive 527 disk0/00/00/10/78 2024-04-05 17:58:58 2024-04-08 12:45:46 2024-04-05 17:58:58 data_collection show Chen Hao hao.chen.cs@gmail.com 0009-0009-5281-4616 Hou Yuqi yxh029@student.bham.ac.uk 0009-0002-3086-8698 Qu Chenyuan cxq134@student.bham.ac.uk 0009-0000-4814-2022 Testini Irene i.testini@bham.ac.uk Hong Xiaohan xxh247@student.bham.ac.uk 0009-0009-1270-8579 Jiao Jianbo j.jiao@bham.ac.uk 0000-0003-0833-5115 10col_ephy Multi-Modal, 360° Panoptic Perspective, Multi-Viewpoint Scene Understanding, Temporal Activity Localisation 360+x dataset introduces a unique panoptic perspective to scene understanding, differentiating itself from existing datasets, by offering multiple viewpoints and modalities, captured from a variety of scenes. Our dataset contains: 1. 2,152 multi-model videos captured by 360° cameras and Spectacles cameras (8,579k frames in total) 2. Capture in 17 cities across 5 countries. 3. Capture in 28 Scenes from Artistic Spaces to Natural Landscapes. 4. Temporal Activity Localisation Labels for 38 action instances for each video. IMPORTANT NOTICE: Due to the large volume of the data files, they have been stored in the BEAR Research Data Store space. If you wish to access the data, please contact research-data@contacts.bham.ac.uk to request and receive the relevant link to the data folder. 2024-04-05 published University of Birmingham https://doi.org/10.25500/edata.bham.00001078 Dataset research-data@contacts.bham.ac.uk School of Computer Science other Ramsay Research Fund Royal Society Short Industry Fellowship Baskerville Tier 2 HPC zh yes yes yes TRUE Assistant Professor Jiao Jianbo j.jiao@bham.ac.uk web https://x360dataset.github.io/ Two main devices were used for our data collection: the Insta 360 One X2 and Snapchat Spectacles 3 cameras. The 360 One X2 has two fish-eye cameras that collect 360◦ panoramic visual information in the scene with 5760*2880 resolution and a frame rate of 25 FPS. Additionally, directional audio was recorded using four microphones in directional audio mode. While the Spectacles 3 has a stereo camera attached to a pair of glasses used to capture the egocentric binocular vision within the scene at a resolution of 2432*1216 and a frame rate of 60 FPS. en 2022 2024