e-Detect: Non-User Mask Detection Based on Image Processing Using Convolutional Neural Network Method

Abstract

The COVID-19 pandemic is a situation that spreads the disease caused by the corona virus. One of the efforts to prevent the spread of the virus from one person to another is by requiring everyone to wear a face mask, especially for those in public areas. There have been several similar previous studies, however, none of them accompanied by follow-up, if someone is found without a face mask. The purpose of this research is to make a detector called e-Detect to detect visitors in public areas such as supermarkets, hospitals, schools and other similar places that without wearing a face mask (non-user face mask) uses the Convolutional Neural Network (CNN) method. If e-Detect detects non-user face mask who will enter through the gate of a public area without wearing face mask, the gate will not open, the buzzer will sound, and the visitor's photo will be sent to the security guard via telegram as a notification. The gate is only open when visitors wear face masks. Experiments have been carried out using 17 types of masks with percentages of: accuracy 94%, precision 100%, sensitivity 94.11%, specificity 100%, and error rate is 5.56%. A trial on e-Detect ability, show in the experiment range detection distance using any type of mask, which is 175 centimeters. All visitors who come to public areas through the gate that has been installed e-Detect can be ensured that the visitor's face will not be more than 175 cm apart, thus all visitors can be ensured to be well supervised. Based on this data, it can be said that e-Detect is feasible to be produced and used as an effort to prevent the spread of COVID-19.