AI-Based Typing Biometric for Behavior-Locked Decryption Using Kernel Ridge Regression

Abstract

Modern information systems require strong security mechanisms to protect sensitive data from unauthorized access. Traditional authentication methods such as passwords and PINs are widely used, but they are vulnerable to security threats like password theft, phishing attacks, and credential leakage. In many cases, attackers can gain access to confidential data simply by obtaining the correct password. Therefore, there is an increasing need for intelligent authentication systems that can verify the identity of users more accurately and securely. This paper proposes an AI-Based Typing Biometric System for Behaviour-Locked Decryption using Kernel Ridge Regression with Face Recognition to enhance system security. The proposed system utilizes behavioural biometrics by analysing a user's typing pattern, also known as keystroke dynamics. Typing features such as dwell time, flight time, and typing rhythm are captured when the user enters credentials. These features are processed using the Kernel Ridge Regression (KRR) machine learning algorithm to identify unique typing behaviour patterns of authorized users. In addition to typing biometrics, the system integrates a face recognition module as a second-level authentication mechanism. The system captures the user's facial image through a camera and compares it with stored facial data to verify the user’s identity. Only when both the typing pattern and facial recognition match the registered user profile will the system allow the behaviour-locked decryption of protected data .In this context, the proposed project introduces an AI-Based By combining behavioural and physiological biometrics, the proposed system provides a multi-factor authentication framework that significantly improves data security and reduces the risk of unauthorized access. The system can be effectively used in applications such as secure login systems, financial platforms, and confidential data protection environments.