Efficient Face Recognition System for Education Sectors in COVID-19 Pandemic

COVID-19 has completely startled the education sectors. Educational organisations are seeking technological innovations to execute their academic activities smoothly e.g. online classes, exams, meetings etc. Face recognition system is one of the main technological innovations that is helping the education sectors to verify the student's identification. In this paper, the authors aimed to explore and develop an efficient face recognition system. These systems require enormous visual information. Visual information holds a large set of features that needs large spatial occupancy. Convolutional neural networks (CNNs) are the most suitable techniques to classify large spatial datasets. We proposed an experimentally based 8 layered tuned CNN model comprising four convolution layers, three max-pooling layers, and one dense layer. The proposed model is trained on 14 face classes and analysed the recognition performance based on different image resolutions, activation functions, batch sizes and optimisers. The proposed model performance has been compared to the state-of-the-art methods and achieved high face recognition.