Memories Upload 推出第一视角视频平台，为机器人训练提供逐帧手部标注数据
做机器人的朋友可以看看这个新数据平台：第一视角视频逐帧标注每只手的动作，能按动作检索片段直接拿来训练。相比自己搭遥操作采集，省了不少成本，每月还在上新数据。
Memories Upload 平台正式推出，定位为面向机器人训练的数据采集网络，将日常工作的第一视角视频转化为训练数据。平台对每段视频按帧进行手部动作标注，机器人实验室可以按每只手正在执行的动作进行检索，只挑选需要的片段用于训练。该网络每月新增视频素材，为实验室提供持续增长的真实世界数据来源，缓解机器人领域训练数据稀缺的问题。
机器人训练长期依赖人工遥操作或仿真数据，成本高且难以规模化。逐帧手部标注的思路让实验室能像搜索数据库一样筛选动作片段，把「徒手做任务」的视频变成可检索的监督数据。这类第一视角采集网络如果持续扩充，可能改变机器人实验室获取真实操作数据的方式，降低数据获取门槛。
https://t.co/ytdbGDhE18 has introduced Memories Upload, a capture network that transforms first-person video of everyday work into training data for robots.
The platform labels footage frame by frame for each hand, letting robotics labs search by what each hand is doing and select only the moments they need for training.
https://t.co/ytdbGDhE18 says the network adds new footage every month, giving labs a growing supply of real-world data.
This report is categorized under AITOP's "Products" desk. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include 机器人训练, Memories Upload, 数据标注, 第一视角视频.
Source attribution: curated from TestingCatalog (@testingcatalog). Cross-source corroboration: 2 independent sources have covered this development (including AI Breakfast), confirming industry-wide signal strength.
- AI Breakfast15:41Source
