The objective of this project is to design and develop an AI-based secure login system that uses behavioral biometrics such as typing patterns and mouse movements for user authentication. The system enhances security by continuously verifying user identity using machine learning techniques.
Study traditional authentication methods and limitations in password-based systems.
Analyze behavioral biometric concepts such as keystroke dynamics, typing speed, mouse movement patterns, and login timing behavior.
Prepare Software Requirement Specification (SRS) and system architecture documentation.
Design system architecture integrating login module, behavioral data capture module, and AI-based classification engine.
Create database schema for users, behavioral data samples, authentication logs, and trust scores.
Implement secure user registration with password hashing.
Develop module to capture behavioral data during login attempts.
Preprocess collected behavioral data for model training.
Implement a simple machine learning model (e.g., Logistic Regression / Decision Tree for BCA; Neural Network for MCA).
Train the model to recognize legitimate user behavior patterns.
Calculate trust score during login and trigger additional verification if anomaly is detected.
Implement fallback authentication (OTP/PIN) when behavioral mismatch occurs.
Maintain logs of login attempts and anomaly detection results.
Apply input validation and secure coding practices.
Perform model evaluation using accuracy and confusion matrix metrics.
Conduct unit testing and system testing.
Prepare documentation including architecture diagrams, ML workflow, ER diagrams, and test cases.
Deploy system locally for demonstration and analysis.
