Exploring the Spectrum of Biometric Technologies: A Systematic Literature Review of Conventional and Unconventional Modalities
W.P. Aldo Arista, K.S. Bryan Fernando, Sidharta Sidharta · Procedia Computer Science · 2024
Biometric recognition is a process of validating and identifying people based on their distinct physiological or behavioral traits. Through this literature review, we strive to reduce the knowledge gap among the public about existing biometric modalities, how the biometrics technology is applied, what algorithms are often used in processing the biometrics data, and what common benchmarks are often used to test and measure the performance of each biometrics. This paper provides a systematic literature review with the Kitchenham and Cochrane method, aimed at exploring variant types of biometric modalities and how each of them collects data, how biometric modalities are applied in various industries, common algorithms or methods used in biometric technologies, and establishing metrics and benchmarks for evaluating the performance of biometric technology. As a result, each biometric modality has its way of collecting data, for example, fingerprint recognition uses techniques like contrast correction, noise filtering, and sharpening to extract the fingerprint pattern. Biometric modalities are already used in many sectors of industries, and fingerprint readers or facial feature detection cameras are used to reduce the authentication process in the banking sector. Minutiae-based algorithm commonly used for fingerprint biometric modality, Eigenface for facial detection, and Euclidean distance or Support Vector Machine (SVM) for gait detection. Accuracy, False Acceptance Rate (FAR), False Rejection Rate (FRR), Equal Error Rate (EER), and Receiver Operating Characteristic (ROC) metrics and benchmarks are used to evaluate the performance of biometric technology.