Multilevel Face Mask Detection System using Ensemble based Convolution Neural Network

Abstract

Due to COVID outbreak face mask detection in the industry as well as in any gathering playing an important role. Either person worn facemask or not worn. In this CORONA situation, Use of a face mask is one such preventative that is crucial. Facial recognition technologies are now used by many businesses and organizations for their own general purposes. We are all aware of how important it has become to always wear a mask when we travel. However, as we all know, it is impossible to monitor who is wearing a mask and who is not. If some person who worn the mask, then it is not confirmed whether he/she worn it correctly or not. We make the use of AI in our daily life. We achieve this with the help of a deep learning, where we train the model using various convolution neural network approaches and created a hybrid model using bagging-based ensemble learning. Here, detection is performed based on voting-based classification so that we can enhance the accuracy of our model. We have found dataset from MAFA and Kaggle. The hybrid approach of C2N model achieved exceptional accuracy with the use of a dataset of face mask detection that contains both with and without face mask photographs. In our multilevel facemask detection system at the first level our model will predict whether the person worn facemask or not and at its second level it will predict the correctness of facemask, whether it is worn correct or not.