DEVELOPMENT OF BIOMETRIC FACIAL RECOGNITION SYSTEM

1.1 Background of the Study

Recent developments in computer vision, pattern recognition, and machine learning have accelerated the evolution of biometric recognition systems. Due to its non-intrusive nature and context-adaptability, facial recognition technology (FRT) has become a popular tool among other biometric approaches for both security and identity applications (Phillips et al., 2000). FRT has become crucial for organisations and governments looking for effective, automated identity verification solutions as global security issues, particularly in areas like border control, crime prevention, and personal device security, have increased (Wayman et al., 2005). In contrast to fingerprint or iris recognition, facial recognition technology allows for contactless and instantaneous identification verification, which makes it appropriate for high-traffic settings where conventional techniques might not be feasible (El-Abed & Charrier, 2022).

FRT's expansion and enhanced capabilities have been facilitated by the recent rise in internet-connected devices and the incorporation of artificial intelligence (AI) across multiple industries (Pisani et al., 2019). Advanced algorithms power the system's operation by analysing and contrasting stored data in a database with face traits including the distance between the lips, nose, and eyes. These face features, which are frequently referred to as "biometric templates," are specific to each person and enable highly accurate identification (de Luis-Garcı́a et al., 2003). Important advances in artificial intelligence (AI), specifically in deep learning and convolutional neural networks (CNNs), have greatly improved facial recognition systems, allowing them to process images in difficult-to-recognize situations such as dim lighting, partial occlusions, and a range of facial expressions (Jain, Ross, & Prabhakar, 2004).

As FRT grows, its uses are observed in a variety of fields. FRT is becoming a vital tool in law enforcement for solving crimes using surveillance footage or identifying people on watchlists (Collins et al., 2021). Facial recognition is being included into patient identification systems in the healthcare industry to lower the possibility of misidentification and guarantee that patients receive the right care (Yaacoub et al., 2022). Similar to this, FRT is being used more and more in business settings in access control systems to enhance workplace security and secure restricted areas (Wamba-Taguimdje et al., 2020). But in spite of these developments, FRT has also brought up serious privacy, legal, and ethical issues, especially in relation to data security, monitoring, and possible biases in the algorithms (Nguyen et al., 2017).

The possibility of racial and gender biases in FRT systems is a serious problem because research has shown that some demographic groups—like women and ethnic minorities—have higher identification mistake rates than others (Publicover & Marggraff, 2017). These prejudices can result in incorrect identification and have significant ramifications in social and legal situations, igniting continuous discussions about the accountability and fairness of the technology. To address these problems, scholars and decision-makers support the creation of open, objective algorithms and uniform legal frameworks to control the application of FRT (Ometov et al., 2018). In order to protect private rights, a number of countries, especially those in the European Union, have implemented strict regulations restricting the use of FRT (Jain, Ross, & Pankanti, 2006).

Research into FRT's design, ethical considerations, and optimisation for a range of applications is necessary given the quickening pace of technical advancements and its expanding use. Researchers and developers alike are now focused on creating facial recognition systems that are not only efficient but also equitable and socially conscious (Nguyen et al., 2017). In order to add to the expanding corpus of research and useful solutions in this area, this study investigates the architecture and deployment of a biometric facial recognition system, concentrating on resolving typical technical issues including accuracy, security, and bias reduction.

1.2 Statement of the Problem

Even while facial recognition technology has many uses and has advanced, there are still a number of problems that prevent its equitable and broad implementation. Current systems' vulnerability to biases based on gender, race, and other demographic characteristics is a major issue that leads to discrepancies in accuracy among varied groups (Minaee et al., 2023). If left unchecked, these biases can have detrimental effects, such as false identifications, invasions of privacy, and societal prejudice, particularly in delicate domains like law enforcement (Jung et al., 2020). Furthermore, because facial recognition systems are becoming more susceptible to cyberattacks and identity spoofing attempts, the security of biometric data continues to be a major problem (Rathgeb & Uhl, 2011).

Technical difficulties including the requirement for a lot of computing power, environmental constraints like illumination, and managing real-time data streams present serious barriers to system performance and dependability in addition to these ethical and security concerns (Yang et al., 2015). By creating a reliable and secure biometric facial recognition system that reduces bias, protects data, and adjusts to different ambient circumstances, this work seeks to address these issues and increase overall efficacy and user confidence in the technology.

1.3 Objectives of the Study

The objectives of this study are as follows:

1. To design a facial recognition system with high accuracy and minimal bias across diverse demographic groups.

2. To implement security protocols that safeguard biometric data from potential breaches and spoofing attacks.

3. To evaluate the performance of the developed system under various environmental and operational conditions.

1.4 Research Questions

The study seeks to answer the following research questions:

1. How can a facial recognition system be designed to achieve accuracy and reduce demographic biases?

2. What security measures can be integrated to protect biometric data in a facial recognition system?

3. How does the system perform under different environmental conditions, and what improvements can be made?

1.5 Significance of the Study

By addressing significant shortcomings in the state-of-the-art facial recognition technologies, this study adds to the body of knowledge on biometric security. In order to create a system that not only satisfies functional objectives but also adheres to ethical norms, our research will concentrate on minimising bias, improving security, and optimising efficiency. Fair and secure identity verification systems are crucial in a number of industries, including corporate security, healthcare, and law enforcement, where this work may have ramifications (Odelu, Das, & Goswami, 2015). Additionally, given the heightened scrutiny of biometric technology, the results of this study may help direct future research in the creation of more inclusive and privacy-respecting biometric systems (Nguyen & Su, 2023).

1.6 Scope of the Study

This study's scope includes the creation, testing, and assessment of a biometric facial recognition system. Choosing and honing algorithms, putting security measures in place, and testing the system in different scenarios are all included in this. Although the study will mainly concentrate on technological factors like security and accuracy, it will also take ethical considerations and the elimination of bias in system performance into account. Nevertheless, the report does not thoroughly address policy-related concerns or discuss particular uses of the technology in the public or private sectors.

1.7 Definition of Terms

• Biometric Recognition: The automated identification of individuals based on their unique biological and behavioral characteristics.

• Facial Recognition Technology (FRT): A system that uses algorithms to identify or verify a person’s identity using their facial features.

• Bias: In the context of FRT, bias refers to the tendency of the system to perform differently across various demographic groups, often leading to disparities in identification accuracy.

• Data Security: Measures and protocols put in place to protect digital information, especially sensitive or personal data, from unauthorized access or breaches.

• Spoofing Attack: A method used by attackers to trick the system into recognizing a fake biometric input, such as a photograph, as genuine.