D'Aurizio Nicole, Baldi Tommaso Lisini, Paolocci Gianluca, Prattichizzo Domenico
Department of Information Engineering and MathematicsUniversity of Siena 53100 Siena Italy.
Department of Advanced RoboticsIstituto Italiano di Tecnologia (ADVR) 16163 Genova Italy.
IEEE Access. 2020 Jul 27;8:139033-139043. doi: 10.1109/ACCESS.2020.3012309. eCollection 2020.
The alarming morbidity of COVID-19 has drawn the attention to the social role of hygiene rules, with a particular focus on the importance of limiting face-touch occurrences. To deal with this aspect, we present No Face-Touch, a system able to estimate hand proximity to face and notify the user whenever a face-touch movement is detected. In its complete setup, the system consists of an application running on the smartwatch and a wearable accessory. Its ease of implementation allows this solution to be ready-to-use and large-scale deployable. We developed two gesture detection approaches compatible with sensors embedded in recent smartwatches, i.e. inertial and magnetic sensors. After preliminary tests to tune target gesture parameters, we tested the two approaches and compared their accuracy. The final phase of this project consisted in exploiting the most robust approach in a daily living scenario during a 6-days campaign. Experimental results revealed the effectiveness of the proposed system, demonstrating its impact in reducing the number of face-touches and their duration.
新型冠状病毒肺炎（COVID-19）令人担忧的发病率已引起人们对卫生规则社会作用的关注，尤其关注限制触摸面部行为的重要性。为了应对这一方面，我们推出了“无脸触”系统，该系统能够估计手与面部的接近程度，并在检测到任何触摸面部动作时通知用户。在其完整设置中，该系统由运行在智能手表上的应用程序和一个可穿戴配件组成。其易于实施的特点使得该解决方案可以立即使用并大规模部署。我们开发了两种与近期智能手表中嵌入的传感器兼容的手势检测方法，即惯性传感器和磁传感器。在进行初步测试以调整目标手势参数后，我们对这两种方法进行了测试并比较了它们的准确性。该项目的最后阶段是在为期6天的活动中，在日常生活场景中采用最可靠的方法。实验结果揭示了所提出系统的有效性，证明了其在减少触摸面部次数及其持续时间方面的影响。
