Arduino based Anti-Face-Touching Device and prediction model

Abstract

Due of the fast pace at which COVID-19 may spread through respiratory illness, the terrible condition it was in heightened public tension. The WHO's primary recommendations advised against often touching your face in order to avoid the transmission of viruses through your lips, eyes, and nose. According to research, the typical person was discovered to touch their face about 20 times each hour since it is everyone's unconscious behavior. In order to cope with this, the study suggests a hardware model that recognizes hand motions that are made in the direction of the user's face and alerts them to such movements using both aural and visual sensory feedback modalities. In order to create a model for the prediction of facial touch motions, the study analyses deep learning architectures in more detail. The FaceGuard device, which is a deep learning-based prediction model used to determine whether or not a hand movement would result in face contact, is compared to the accuracy of the suggested hardware model in the paper “FaceGuard: A Wearable System To Avoid Face Touching1.” It alerts the user through vibrotactile, aural, and visual sensory modalities. After investigation, it was discovered that the hardware model had less accuracy than the deep learning model and required shorter time to respond to vibro tactile sensory data.