Systematic Review of Facial Recognition Algorithms and Approaches for Crime Investigations

Abstract

Crime control in human societies has continued to pose significant problems and requires dynamic approaches to be subdued through effective investigation mechanisms. Notable among these approaches is the use of biometric facial recognition which has proven to be ideal, due to its flexible and nonintrusive nature. Mainly, this research conducted a systematic review on algorithms and approaches for facial recognition to aid security operatives in crime investigations. For the first time, the review coined and described three operational environments namely, regulated, unregulated and semi-regulated to which facial recognition is applicable. However, semi-regulated environment is yet to be addressed based on its peculiar characteristics. Subsequently, this study proposed the design of a facial recognition system premised on deep learning and local binary patterns histograms algorithm (LBPH). Certainly, future implementation of the design will help to identify and document known and unknown individuals, thus creating a more efficient and effective approach for crime control.