Keystroke Pattern Authentication Using User Posture Profiles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current user authentication methods, such as password-based systems, are vulnerable to attacks and inconvenient, while bio-information methods require additional devices, and there is a need for a more effective and user-friendly solution, especially for smartphones used in online banking and FinTech.
Innovation Solution
A method and apparatus for training and classifying keystroke patterns based on user posture using keystroke data, which includes time information and sensor information, to enhance user classification success rates by creating unique keystroke profiles for each posture or cluster.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If password-based authentication is used, then ease of operation is improved, but security reliability deteriorates due to vulnerability to attacks and easy guessing
Solution Approach 1:
The authentication process is segmented into multiple independent features: keystroke timing patterns, posture information from sensors, and password input. Each feature is analyzed separately and combined to form a comprehensive authentication decision, making it harder for attackers to compromise the entire system while maintaining user convenience.
Solution Approach 2:
The system changes from static password-only authentication to dynamic multi-parameter authentication by incorporating temporal parameters (keystroke timing, duration, intervals) and spatial parameters (posture, device orientation). These additional parameters increase security without significantly impacting ease of operation.
2Reliability
If bio-information authentication is used, then security reliability is improved, but device complexity increases due to requirement of additional devices
Solution Approach 1:
The patent makes the smartphone itself multi-functional by utilizing its existing sensors (accelerometer, gyroscope, keyboard controller) for authentication purposes. The device performs both its standard communication functions and biometric-like authentication using only its built-in capabilities, eliminating the need for separate authentication devices.
Solution Approach 2:
The system enables the smartphone to authenticate users using its own built-in sensors and resources without requiring external biometric devices. The device serves itself by leveraging its existing hardware capabilities (accelerometer for posture detection, keyboard controller for keystroke analysis) to perform authentication functions.
3Ease of operation
If simple keystroke dynamics are used, then ease of operation is maintained, but measurement precision deteriorates due to inability to account for different user postures
Solution Approach 1:
The system applies local quality by creating posture-specific keystroke profiles. Instead of using a single generic authentication model, it tailors the authentication parameters and thresholds to match each detected posture (holding, typing, viewing), thereby improving measurement precision for each local condition while maintaining overall ease of operation.
Solution Approach 2:
The authentication system becomes dynamic by continuously adapting to the user's posture. The system detects posture changes in real-time and adjusts the authentication parameters accordingly, making the measurement precision responsive to changing conditions rather than static, thereby improving accuracy without adding operational complexity.
Data Source
AI summary
A user classification apparatus and method using a keystroke pattern based on a user posture are provided. A user classification method using a keystroke pattern may include receiving keystroke data from a user, extracting feature information corresponding to a unique pattern of the user using the received keystroke data, comparing the extracted feature information to a pattern of a keystroke profile for each posture of the user pre-generated through training, and determining whether currently input keystroke data is data of a trained user based on a result of the comparing.