WiHead: WiFi-Based Head-Pose Estimation
WiHead: WiFi-Based Head-Pose Estimation
复制标题
WiHead：基于 WiFi 的头部姿势估计
DOI：
10.1007/978-3-031-05431-0_5
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Konomi Shin’ichi
中科院分区：
文献类型：
--
作者：
Liu Yiming;Konomi Shin’ichi
由于COVID-19的影响，人们开始进行在线课程或会议。然而，由于缺乏非语言交流，这使得人们很难有效地相互沟通。虽然网络摄像头可用于在线课程等，出于隐私的原因，人们通常不想打开它们。因此，有必要开发隐私保护的方式，使在线学习和工作环境中的非语言交流。WiFi作为一种传感器，可以用来检测非语言的姿态，如头部姿势，并已越来越受到重视，由于其优点，避免光的影响，非视线监测，隐私保护等，在本文中，我们提出了一种方法，它使用WiFi CSI数据来估计头部姿势。该方法不仅利用了原始CSI数据的幅值和相位信息，而且还利用了频域信息。我们的实验与所提出的方法证实了头部姿态估计的WiFi CSI数据的基础上的可行性。这对于无设备感测检测具有重要意义。特别是在网络会议和在线课程广泛使用的当今世界，基于WiFi的头部识别可以在保护隐私的同时向对方提供反馈，有助于提高沟通的质量和舒适度。
Due to the impact of Covid-19, people have started to conduct online courses or meetings. However, this makes it difficult to communicate with each other effectively because of the lack of non-verbal communication. Although webcams are available for online courses, etc., people often do not want to turn them on for privacy reasons. Thus, there is a need to develop privacy preserving way to enable non-verbal communication in online learning and work environments. WiFi as a sensor can be used to detect non-verbal gestures such as head poses, and has been increasingly valued due to its advantages of avoiding the effects of light, non-line of sight monitoring, privacy protection, etc. In this paper, we proposed an approach, which uses WiFi CSI data to estimate head pose. Our approach not only use the amplitude and phase data of raw CSI data, but also use the information in frequency domain. Our experiment with proposed approach confirmed the feasibility of head pose estimation based on WiFi CSI data. This has important implications for device-free sensing detection. Especially in today’s world where web conferences and online courses are widely used, WiFi-based head recognition can give feedback to the other party while protecting privacy, which helps to improve the quality and comfort of communication.