Widar3

该数据集名为Widar3，是一个用于人体活动识别的无线感知数据集，它通过商用Wi-Fi网卡以接收信号强度指示（RSSI）和信道状态信息（CSI）的形式收集而成。该数据集包含了258,000个手势活动的实例，总时长达到8,620分钟，涵盖了75个不同的领域。Widar3数据集持续发展，不断包含更多类型的人体活动，并已广泛受到研究人员的关注。它是该领域内规模最大、内容最全面的数据库。其研究任务主要集中在人体活动识别。

The dataset named Widar3 is a wireless sensing dataset dedicated to human activity recognition. It was acquired using commercial Wi-Fi network interface cards, with data available in the forms of Received Signal Strength Indication (RSSI) and Channel State Information (CSI). The dataset contains 258,000 instances of gesture activities, with a total duration of 8,620 minutes, covering 75 distinct domains. Widar3 is an evolving dataset that continuously adds more types of human activities, and has garnered widespread attention from researchers. It is the largest and most comprehensive database in this field, with research tasks based on it mainly focusing on human activity recognition.

搜集汇总

背景与挑战

背景概述

Widar3数据集是一个基于Wi-Fi的跨领域手势识别数据集，提供了CSI、DFS和BVP等多种特征文件格式，支持无线感知和手势识别研究。该数据集与多个研究项目相关，并已在多个顶级会议和期刊上发表。