A Novel Automated Human Face Recognition and Temperature Detection System Using Deep Neural Networks—FRTDS

This paper proposes a novel FRTDS (Face Recognition and Temperature Detection System) that is contactless and performs real-time face recognition. The system had proved to be fast, built at low-cost, and efficient in user authentication. FRTDS consists of numerous algorithms and techniques that have been implemented to improve the performance of the entire system with the help of Deep Neural Networks. FRTDS can capture images from a video stream and can detect faces from 70–90 cm away from the camera. An interactive front-end recognizes and displays the identity of the person. FRTDS also includes a temperature sensor to monitor the health of the person, before they enter any premises. The recognized face along with temperature data is stored at the back-end with the current time and date. This paper also presents a novel customized tool that eases the process of dataset creation and augmentation, and a novel prediction discrepancy elimination algorithm. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.