Human Activity Recognition
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/srvcodes/continual-learning-benchmark/tree/master/datasets
下载链接
链接失效反馈官方服务：
资源简介：
该数据集主要关注使用加速度计、陀螺仪和身体传感器读数进行人体活动识别的任务。在五个跨用户实验中，模型在一个用户的数据上进行训练，并在不同用户的数据上进行测试。每个样本包含9个通道和128个数据点，旨在完成人体活动识别的任务。
This dataset primarily focuses on the task of Human Activity Recognition (HAR) using readings from accelerometers, gyroscopes, and body-worn sensors. In five cross-user experiments, models are trained on data from one single user and tested on data from distinct users. Each sample comprises 9 channels and 128 data points, and is tailored for the human activity recognition task.
