emg2qwerty
收藏资源简介:
emg2qwerty数据集是由Meta公司创建的一个大规模非侵入性肌电图信号数据集,专门用于触摸打字任务。该数据集包含108名用户在打字时手腕处记录的1,135个会话,总计346小时的数据,是目前最大的公开sEMG数据集。数据集的创建过程包括精确的地面实况注释和可重复的基线模型。该数据集的应用领域主要集中在提高人机交互的带宽,特别是在增强现实和虚拟现实环境中,旨在解决现有输入方式(如键盘、鼠标和触摸屏)带宽受限的问题。
The emg2qwerty dataset is a large-scale non-invasive electromyography (EMG) dataset created by Meta, specifically designed for touch typing tasks. It contains 1,135 sessions of wrist EMG signals recorded from 108 users during typing, totaling 346 hours of data, making it the largest publicly available sEMG dataset to date. The dataset creation process includes precise ground truth annotations and reproducible baseline models. Its application areas mainly focus on improving the bandwidth of human-computer interaction, especially in augmented reality (AR) and virtual reality (VR) environments, aiming to address the bandwidth limitation issues of existing input methods such as keyboards, mice and touchscreens.




