giu-alb/something_something_v2

Something Something数据集（版本2）是一个包含220,847个标记视频剪辑的集合，展示了人类使用日常物品执行预定义的基本动作。该数据集旨在训练机器学习模型，以精细理解人类手势，例如将某物放入某物中、将某物倒置以及用某物覆盖某物。数据集支持动作识别任务，所有注释均为英文。数据集的创建目的是通过视频提供关于物理世界的详细知识，以帮助神经网络更深入地理解复杂场景和情境。

The Something-Something v2 dataset is a collection of 220,847 labeled video clips of humans performing pre-defined, basic actions with everyday objects. It is designed to train machine learning models in fine-grained understanding of human hand gestures like putting something into something, turning something upside down and covering something with something. The annotations in the dataset are in English, and it is intended for action recognition tasks. The dataset is curated through crowdsourcing, with annotations provided by AMT workers. The dataset is licensed under a custom license defined by Qualcomm.