A novel coronavirus, named COVID-19 had resulted in an outbreak of viral pneumonia in China and became a pandemic in March 2020. According to WHO, 6.2 million people had affected by the disease till November 2020. Until recent updates, any proven cure has not been discovered yet. However, wearing a facemask and social distancing are predominant preventions. In the current scenario, person identification from masked images and without the use of biometric machines has become a challenge. In this paper, we have generated a dataset of almost 20000 pictures with the Raspberry Pi Camera Module. This dataset was divided into 20 classes, out of which, 10 classes are of masked face images and, 10 are of thumb images without using a biometric machine. In this article, we have proposed two less complex convolutional neural networks (CNN) based classifiers that have accurately classified the data and recognized the people in real-time using a Raspberry Pi 4 module. We achieved an accuracy of 97.67% in masked face recognition and 100% accuracy to recognize the thumb images.