Real time analysis of unmask face detection in human skin using tensor flow package and IoT algorithm

ABSTRACT

In covid19, security provided by Indian government, due to severe virus speed to sudden death. This proposed method solved face detection, mask detection and thermal value monitoring for security purpose. We need amount of testing and need tests 50000/day to measure count of infection. Government produces a report that maintain social distancing, wearing mask when in outside and proper testing. Most people do not serious to wearing mask. We focused the counting of Unmask-persons using opencv and MTCNN algorithm. Faces conditions are changing due to various atmospheric season and light vision. The MTCNN algorithm adjusted video streaming data with Mask detection accuracy at >90%. A database stored in Excel file or send Emergency message to email or buzzer kept on, when unmask visit at any places like that shopping complex, hotel, hospital, Traffic areas and temple. MTCNN- Multi-Task Convolutional Neural Network algorithm solved facial recognition problems. We have proposed the algorithm for Mask detection and Thermals value detection using Pycharm-python and implemented in Raspberry pi4. Support Vector Machine (SVM) supported tasks of MTCNN and stored in database. The outcome of the system is Alert through monitor and SMS to officials Government persons, when capturing unmasked persons visit in live streaming video using Opencv – Python.