ACTIVE（Action from Robotic View）是一个专为移动服务机器人感知视角设计的大规模数据集，包含30个复合动作类别、80名参与者和46,868个注释视频实例，涵盖RGB和点云两种模态。参与者在不同环境中执行各种人类动作，距离从3米到50米不等，同时摄像机平台也是移动的，模拟了机器人感知的真实场景。该数据集旨在推动自然人机交互（N-HRI）中的动作和属性识别研究。

github2025-07-31 更新230

Bangla_Sign_Language

孟加拉手语识别

手势识别

Annotated Dataset for Gesture Recognition

kaggle2024-10-21 更新80

Rock_Paper_Scissors

图像识别

手势识别

displayName: Rock Paper Scissors license: - CC BY 2.0 paperUrl: "" publishDate: "2019" publishUrl: https://laurencemoroney.com/datasets.html#google_vignette publisher: - Laurence Moroney tags: - Gestu

魔搭社区2025-12-04 更新1020

BeyondDeskVR: an extended virtual hand interaction system in virtual reality

虚拟现实

手势识别

As virtual reality (VR) technology advances, the use of VR scenarios and applications is becoming increasingly prevalent, and the fatigue issues associated with existing mid-air hand interactions are

DataCite Commons2025-10-13 更新80

Preprocessed custom dataset

资源简介：

相关数据集