Description of Invention
The project is about a continuous authentication system leveraging touch stroke analysis to enhance security on Android devices. Traditional authentication methods, such as passwords and PINs, are increasingly vulnerable to sophisticated attacks. This system aims to address these vulnerabilities by continuously monitoring the user's touch patterns, such as pressure, speed, and coordinate movement on the device, using a One Class Support Vector Machine (SVM) model. The model identifies atypical user behavior, enabling real-time detection of unauthorized access. Through a novel combination of behavioral biometrics and machine learning, the system enhances security by dynamically adapting to individual user profiles while remaining unobtrusive. The findings indicate the model's effectiveness, detecting 82 anomalies within the dataset and classifying 779 points as normal behavior, demonstrating its potential in safeguarding mobile devices against unauthorized use.