An Improved Machine Learning Approach to Detect Real Time Face Mask

Several nations have implemented health protocols like maintaining a particular measure of distance from each other and use of face masks when going out in public, in an effort to stop or at least reduce the spread of Covid-19. However, manually checking whether each person have put on a mask or not is a tiring job, and is possible only if there is a particular person assigned specially for that. This paves way for the need of an electronic device or a machine that would identify whether a person has worn mask or not. Thus, this research proposes a face mask detection system using a machine learning algorithm known as Support Vector Machine (SVM). After creating and preprocessing the dataset, training the model, and evaluating the final model, an accuracy of 98% has been obtained. The model can further be developed and used in real time scenarios to detect faces without a mask and pass those faces separately into a neural network with the help of CNN to easily find out his/her identity, and punish accordingly.