4/11/2024 On Monday, corresponding to

The master's student (Huda Saddam Mathkal) was discussed in the department council hall about her thesis entitled:

Design and Implementation of multi-factor biometrics identification

The discussion committee consisted of the gentlemen listed below:

Prof..Dr. Muhammad Issam Younis / Department of Computer Engineering - College of Engineering - University of Baghdad. Chairman

Assistant Professor Hadeel Qasim Wadi / Department of Biomedical Engineering - College of Engineering - University of Nahrain. Member

Assistant Professor Atheel Nofal Muhammad Taher / Department of Computer Engineering - College of Engineering - University of Nahrain. Member

Dr. Ahmed Hani Youssef / University of Nahrain / College of Engineering. Member and Supervisor

The student has fulfilled the requirements for obtaining a master's degree in computer engineering

Abstract

In recent years, identity verification systems have suffered from many frauds, resulting in the need for a Substantial immanent security verification process like a biometric system. Biometrics is the technical term for biometric features or characteristics of the body measurements.  The efficiency and effectiveness of security systems depend on the robustness of their associated biometric system in extracting human features efficiently and recognising people with high accuracy. This work presents constructing and implementing a multi-factor biometric system based on three biometric systems (face, iris, and fingerprint), which work separately with five datasets used. In the proposed system, initially, the decisions for the (face and iris) biometrics are obtained after completing the authentication phase for each one separately. In cases where either of these two biometric subsystems fails to recognise the individual, the third biometric identification (fingerprint) is activated to determine whether the individual can be successfully identified. The proposed system employs a number of advanced DL and ML algorithms, where the face identification system uses the YOLOv8 algorithm for feature extraction and classification with head pose images (HPI) and Celeb-A datasets, and utilises the CNN network for feature extraction and an SVM classifier for iris identification system with Multi-Media University (MMU) and Ahmed Myaser Fathi (AMF) datasets. Besides that, CNN algorithm was used for feature extraction and classifying for fingerprint identification using the Sokoto Coventry Fingerprint (SOCOFing) dataset. The implementation results of the multi-factor system show that the proposed algorithms achieve exceptional accuracy on the different datasets, where the proposed system achieved maximum accuracy of up to 99.25%. As a result of the reliability of these biometric features, the results show superior performance, where brilliant results in terms of performance efficiency and high accuracy were was achieved after using different techniques and diverse datasets