← BACK_TO_BASE
PROJECT_01 · CASE FILE OPEN
Digital Behavioural Authentication System
Passwords alone are no longer enough. This system continuously verifies user identity by analyzing behavioral biometrics — how you type, move, and interact — instead of relying solely on a login credential.
PythonFlaskSQLitescikit-learnHTMLCSSJavaScript
/threat
Problem Statement
Traditional authentication only verifies users at login. If an attacker takes over a session, the system cannot distinguish them from the legitimate user. Continuous behavioral authentication closes that gap.
/mission
Objective
Develop an AI-powered behavioral authentication system that continuously validates user identity throughout an active session.
/pipeline
System Architecture
- 01→User Registration
- 02→Login Authentication (Username + Password + Session)
- 03→Behavioral Data Collection (Keyboard + Mouse)
- 04→Feature Extraction
- 05→ML Model Prediction
- 06→Continuous Session Validation
- 07→Alert / Reauth / Logout on Anomaly
/capabilities
Core Features
Behavioral Signals
Keyboard DynamicsMouse MovementTyping SpeedTyping RhythmClick FrequencyCursor VelocityIdle TimeSession Duration
Feature Extraction
Avg typing speedAvg keystroke intervalMouse accelerationMouse pathReaction time
ML Pipeline
CollectCleanNormalizeTrainPredictAuthenticate
Continuous Monitoring
Post-login validationAnomaly detectionAlert generationForced reauthentication
/next
Future Roadmap
Deep Learning (Keystroke CNN)Touch BiometricsFace RecognitionVoice BiometricsMobile AuthenticationCloud DeploymentExplainable AIAdaptive Learning