ABSTRACT
The world communities have suffered from the COVID-19 pandemic for the last two years. Even though many countries have started to normalise the situation, the COVID-19 still becomes a severe threat in the future. Healthy habits, such as complete and frequent handwashing, still need to be practised. These habits can minimise the transmission risks. The paper proposed a single-board computer system that aims to assess the handwashing steps. The standardised handwashing procedure is used to validate the acquired video of hand movement. The system is installed in a Raspberry Pi and receives video data from the connected mini camera. The deep learning model is implemented to provide classification capabilities. The assessment result is summarised according to the movement completeness and total duration. The testing stages found that the proposed system can provide accuracy and F1-score values of 82.55% and 86.66%, respectively. © 2021 ACM.
ABSTRACT
The pandemic situation of COVID-19 is still ongoing in every part of the world countries. Up to now, the medicine and the vaccination for curing the COVID-19 are not yet available. Since the vaccine and medication of COVID-19 are not yet available, the government in many countries are strongly socialising three kinds of new practices. The new habits are physical distancing, wearing a mask, and handwashing frequently. This paper focuses on assessing the completeness of handwashing. The handwashing frames are extracted from a video clip. The frames are then categorised into six-movement steps by using a deep learning-based algorithm. The frame of hand object is separated from the video by applying skin colour differentiation. Two kinds of experiments based on two different video sources are performed in the paper. Scoring simulation is also conducted in the research. The results show that the proposed method can give an excellent performance with accuracy values not less than 90%.