Two Person Interaction Kinect Dataset

人体活动识别具有影响从监视到人机界面再到基于内容的视频检索等广泛应用的潜力。最近，廉价的深度传感器（例如Microsoft Kinect）的迅速发展为活动识别应用程序的实时全身人体跟踪提供了足够的精度。Human activity recognition holds the potential to impact a wide range of applications, spanning from surveillance, human-computer interaction to content-based video retrieval. Recently, the rapid advancement of low-cost depth sensors (e.g., Microsoft Kinect) has provided sufficient accuracy for real-time full-body human tracking in activity recognition applications.

搜集汇总

背景与挑战

背景概述

该数据集专注于两人互动的人体活动识别，利用Microsoft Kinect深度传感器采集同步的视频、深度和运动捕捉数据。它旨在支持活动识别应用的研究，相关成果已在CVPR 2012的研讨会上发表。