Unverified50% confidenceFactExact time
The research team built a dataset containing 30,000 hours of first-person video, covering more than 1,000 scene types, shot by 14,000 contributors.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedEgo4D项目汇集了全球13个大学和研究实验室的3670小时第一人称视频，是目前最大的以自我为中心的视频数据集之一70% similarUnverifiedImageNet contains 14 million images and required tens of thousands of hours of manual annotation67% similarUnverifiedImageNet was led by Professor Fei-Fei Li beginning in 2007 and contains over 14 million manually annotated images covering more than 20,000 object categories.66% similarUnverifiedWhen a multimodal model analyzes a complete webpage screenshot, the number of visual tokens it processes can reach as many as thousands or even tens of thousands.64% similarUnverifiedImageNet has 14 million annotated images, and COCO has 330,000 annotated images63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1012793API
curl https://kongchang.com/api/v1/knowledge/claims/1012793MCP
get_claim(id=1012793)