Face Mask Detection: A Deep Learning Concept

Face Mask Detection program developed by OpenCV, Keras / Tensor Flow uses Deep Learning ideas and laptop Vision to detect face masks on continuous still images such as video streaming. Within the gift situation thanks to COVID-19, there are no requests for the acquisition of a mask to save the area that is currently in dire need of transport, crowded areas, accommodations, major manufacturers and various businesses to guarantee the safety. Also, lack of big image data with a mask has made this task even more difficult and difficult. Following the outbreak of the global COVID-19 epidemic, there was an urgent need for preventive measures, with a mask on the face. The primary purpose of the project is to determine whether there is a face mask on people’s faces in video and photos streamed live. We used the depth of learning to build our face detection model. The features used for object detection are the Single shot detector (SSD) due to its performance with precision and high speed. Apart from this, we used basic concepts to transfer learning to neural networks without ultimately excluding the presence or absence of facial image in a photo or video stream. Test results show that our model works at 100% efficiency with 99% accuracy of test and memory, respectively. Our mask detector did not use any data from the inserted images. The model is accurate, and as we are accustomed to using the design of MobileNetV2, savings are calculated collaboratively and therefore make it a lot easier to move the model to embedded programs like Google Coral & Raspberry Pi etc. This program will be used for time programs that require the acquisition of face masks for security operations due to the emergence of COVID-19. The project is used with the installed plans for use at airports, train stations, offices, school premises and marketplaces to ensure that the local unit of community safety tips is followed.