Anomaly Detection in Biometric Authentication Dataset Using Recurrent Neural Networks
R. Mary Sophia Chitra, Anusha Bamini A M, Chenthil Jegan TM, K. Padmaveni · Advances in digital crime, forensics, and cyber terrorism book series · 2022
In the biometric authentication, the stored data is used for the verification of used identity. The unique biological traits commonly used for biological authentication are facial characteristics, fingerprints, and retinas. It also offers superior fraud detection and customer satisfaction, compared to all other traditional multi factor authentication. Deep learning algorithms plays a major role in anomaly detection and fraud identification in various real-time applications. RNNs have proven that they work well in analysing and detecting anomalies in time series data. RNNs have the unique ability for each cell to have its own memory of all the previous cells before it. This allows for RNNs to process sequential data in time steps which other machine learning models cannot do. RNNs can also be found sorting through your emails to sort out spam and phishing emails from friendly emails. This chapter reviews the methodologies, purposes, results, and the benefits of RNNs in anomaly detection in biometric authentication.