| Title |
Deep Learning-based Thermal Image Mask Recognition for Mobile Environments and Classification using A Cross-entropy Loss Function |
| Authors |
문종원(Moon) ; 윤호섭(Jong-Won) |
| DOI |
https://doi.org/10.5573/ieie.2022.59.1.118 |
| Keywords |
Thermal image; Mask detection; Mask recognition; Mobile environments; Android |
| Abstract |
COVID-19, a global infectious disease, has increased the demand for contactless thermal imaging systems to detect some people with high heat in the crowd. In public places, mask-wearing is mandatory and access is restricted with management personnel at each entrance, to control smoothly with a minimum number of people, detecting masks from RGB images generally has more disadvantages than detecting them from thermal images. In addition to the advantages of using thermal images, this paper introduces deep learning-based thermal image mask detection based on using a lightweight deep learning model that can be sufficiently used in mobile environments. A very lightweight model is used for real-time use even in ordinary mobile phones, but the problem of mask detection is a relatively simple binary classification problem. |