Keystroke Dynamics Authentication Using Derived Data Values
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing keystroke dynamics authentication systems face challenges in accurately distinguishing authorized users from impostors due to high false acceptance and rejection rates, particularly when relying on raw keystroke timing data without advanced processing.
Innovation Solution
The system computes first-order and second-order derived data values from raw keystroke timing measurements, such as dwell times and flight times, to create a template that captures a user's typing style, which is then used for authentication, reducing false acceptance and rejection rates by analyzing dwell tendency and other metrics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If simple statistical methods with Euclidean distance are used for keystroke authentication, then the system is easy to implement, but the false acceptance and rejection rates are high
Solution Approach 1:
The patent transforms raw keystroke timing data into multiple derived parameters including first-order derivatives (dwell time, flight time, typing speed) and second-order derivatives (acceleration, deceleration, dwell tendency). This parameter transformation resolves the contradiction by creating a more comprehensive feature set that improves authentication accuracy while maintaining computational feasibility through systematic data processing.
Solution Approach 2:
The patent extends the authentication approach from simple timing measurements to multi-dimensional analysis by incorporating first-order and second-order derived data. This dimensional expansion captures more aspects of typing behavior (temporal patterns, acceleration patterns, dwell tendencies) thereby improving reliability without sacrificing implementability.
2Reliability
If more complex analysis methods like neural networks are used, then authentication accuracy improves, but system complexity increases
Solution Approach 1:
The patent segments the authentication process into distinct stages: data collection, first-order derivative computation, second-order derivative computation, and template comparison. This segmentation allows complex analysis to be broken down into manageable computational steps, improving accuracy while controlling system complexity through structured processing.
Solution Approach 2:
The patent performs preliminary computations of derived parameters during the enrollment phase, creating templates that capture the user's typing characteristics. This preliminary action separates the complex analysis work from the real-time authentication process, improving accuracy while maintaining operational simplicity during actual authentication.
Data Source
AI summary
A keystroke dynamics authentication system collects measurements as a user types a phrase on a keyboard. A first set of derived data values are computed based on the collected measurements, and then a second set of derived data values are computed based on the first set of derived values. The first and second sets of derived values are used to construct a template for identifying the user based on his typing.