Authors : Tamara Regina; Christopher Matthew

Volume/Issue : Volume 8 - 2023, Issue 6 - June

Google Scholar : https://bit.ly/3TmGbDi

Scribd : https://tinyurl.com/yzrwek3y

DOI : https://doi.org/10.5281/zenodo.8126042

Abstract : The development of effective surveillance and security systems is crucial for public safety and for preventing criminal activities. However, lack of proper infrastructure and technological advancements has hindered the progress of surveillance and security systems in many parts of the world. One significant challenge in this regard is the difficulty in suspect identification from manually inspecting several footags. In this paper, we propose the optimization of facial recognition systems, commonly used in surveillance and safety, through several image processing techniques such as Eigen Faces, Fisher Faces, and Facial Coordinates based on the minimum cross entropy theory. Our approach involves several features incorporating facial recognition that can be analyzed to identify potential suspects. By the end of this paper, we found an interesting result that only one can accurately predict facial recognition from video footage given the training image.