X-RAGE
收藏资源简介:
X-RAGE数据集是由印度理工学院德里分校和Meta Reality Labs Research合作创建的,专门用于扩展现实(XR)设备的手势和动作识别。该数据集利用事件相机捕捉第一人称视角的手势数据,包含36种独特的手势,涵盖单手和双手动作,总数据量为8064条。数据集在多种环境(包括室内和室外)和不同光照条件下录制,以模拟真实世界的应用场景。创建过程中考虑了环境光照和个性化差异,旨在提供一个全面的基准,适用于实际部署。该数据集主要应用于XR设备的手势识别,旨在解决传统基于帧的视觉系统在数据带宽和快速运动捕捉方面的局限性。
The X-RAGE Dataset was co-developed by the Indian Institute of Technology Delhi and Meta Reality Labs Research, specifically tailored for gesture and motion recognition on extended reality (XR) devices. It utilizes event cameras to capture first-person perspective gesture data, encompassing 36 distinct gestures covering both single-handed and two-handed movements, with a total of 8064 data samples. The dataset was collected across diverse environments including indoor and outdoor settings, and under varying lighting conditions to simulate real-world application scenarios. During its development, considerations were given to ambient lighting variations and individual differences, aiming to provide a comprehensive benchmark suitable for practical real-world deployments. This dataset is primarily applied to gesture recognition for XR devices, with the goal of addressing the limitations of traditional frame-based vision systems in terms of data bandwidth and fast motion capture.

- 1印度理工学院德里分校、Meta Reality Labs Research · 2024年



