Thermal Face Recognition: A Comprehensive Review Of Databases And Methodologies

Face identification is the task of recognizing a human face using a digital system, typically for security purposes. Because the face is such an important feature of the human body, it can be used to individually identify a person. For various applications, concentrating the face for verification or identification is critical. Despite recent breakthroughs in face recognition systems, they continue to have major issues due to the wide range of variations in the person’s face (changes like glasses, head tilt, light, and different emotional modes). Each of these issues has the potential to drastically affect the accuracy of face recognition. This research examines the performance of thermal face detection approaches that were described in the existing literature within last seven years (2016–2022). A comparative table of the works covered is also included. The comparative analysis given PUCV-VTF face database that proposed StyleCLIP, GAN structure achieves a higher detection rate compared with other traditional detection such as DTFA, LBP and moments invariant. Nonetheless, using thermal imaging as an effective and distinctive option has gained greater attention among the proposed options in recent years.