The Covid-19 pandemic has facilitated many changes in our day-to-day life including working from home, distance learning, reduced capacity in malls and other indoor places. While restrictions are now being lifted in some countries, social distancing is still practiced. As countries start to restart economic activities, schools and universities start to go back to in-person learning, it is important that another pandemic is avoided. This paper discusses and presents a methodology for one of the ways artificial intelligence can be used to aid in the detection of Covid-19 prevention measures that are implemented to negate the effect of the pandemic. Use of such measures can ensure that local authorities can enforce these measures without being in risk themselves. The proposed method uses a MobileNetv2 model pre trained using the ImageNet dataset as a base model and a FC head layer is fine-tuned onto it to achieve fast and accurate real time detection. The model is trained using two different datasets; one small and one big to see the effects of the size of the dataset on the accuracy of detec-tion. The first stage detects all the faces in the frame after which the mask detection model predicts whether a mask is worn or not. If a mask is worn incorrectly, no mask is predicted. If a violation is detected, an email alert is sent to notify the authorities. After testing, a highly accurate model is obtained which requires low computational power and can be run in real time.


This content is not covered by the Open Government Licence. Please see source record or item for information on rights and permissions.

Cite as

Parkar, B., Soori, P. & Shetty, P. 2024, 'Face Mask Detection and Alert System Using Artificial Intelligence for Covid-19 Prevention', 19th Control Instrumentation System Conference 2022, pp. 231-240. https://doi.org/10.1007/978-981-99-9554-7_18

Downloadable citations

Download HTML citationHTML Download BIB citationBIB Download RIS citationRIS
Last updated: 07 June 2024
Was this page helpful?