Forensic Technology Based on Biometrics Application

Abstract

Human Biometrics have been useful in identifying suspects in forensic investigations, notably in digital forensics. The identification of a person is verified using their biometrics. Because the traditional techniques for a suspect's identification in forensics are no longer accurate and have several flaws, digitalisation was utilised to recognise suspects using images of their biometrics traits. Various techniques were used to recognise and extract features from these images, including continuous and discrete forms of Momentum, algebraic filters, neural networks, and deep learning. Several biometrics have been utilised for this purpose, the most prominent of which we will use in this study are (Face and iris). This is based on databases that include images of multiple people's biometrics. The features of these images are extracted following pre-processing using many feature extractors. This article described a model for a system that stores two types of biometric images for each individual, compares each image to its own dataset, and then recognises and extracts the features of each image. Applied algorithms to extract unique traits for each biometrics used in the search: Face and iris, used a hybrid algorithm of Zernike moment and SVD to identify the facial features and a hybrid algorithm of Legendre moment and LQP to recognise the iris features. The suggested approaches were used in several databases, including ORL and Brazilian for the face, which yielded a rate of 100% and 98% , and CASIA-IrisV4-Interval datasets for the iris, which yielded a rate of 97.29 %.