Typing Cadence Authentication via Statistical Motion Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing user authentication methods through keyboarding habits face challenges in reliably comparing typing samples due to the uniqueness and variability of human muscle movements, leading to unreliable identity verification.
Innovation Solution
The system captures and compares the cadence and habit of users' motions during input device usage, utilizing statistical analysis and geometric constructs to authenticate users by matching current typing samples with a cluster of previously accumulated strings in a user's profile, accounting for differences in timing and muscle movements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional keyboarding habit authentication is used, then user authentication is provided, but reliability is low due to variability in human muscle movements
Solution Approach 1:
The patent segments the typing authentication process into distinct phases: enrollment phase where multiple typing samples are collected and stored in a cluster, and verification phase where new typing samples are compared against the stored cluster. This segmentation allows the system to account for natural variability by comparing against multiple reference samples rather than a single rigid pattern.
Solution Approach 2:
The system performs preliminary action by collecting and storing multiple typing samples during an enrollment phase before actual authentication occurs. These pre-collected samples form a reference cluster that captures the user's natural typing variability, enabling more reliable comparison during subsequent authentication attempts.
2Measurement precision
If strict comparison of typing samples is performed, then measurement precision is improved, but reliability decreases due to natural variability in muscle movements
Solution Approach 1:
The patent implements dynamics by creating a flexible comparison system that adapts to natural typing variability. Instead of using fixed, rigid thresholds, the system dynamically evaluates whether new typing samples fall within the established cluster of the user's typical typing patterns, allowing for natural variations while maintaining security.
Solution Approach 2:
The system changes parameters by transitioning from comparing single typing samples against a single reference to comparing samples against a multi-dimensional cluster of reference samples. This parameter change allows the system to accommodate natural variability in typing patterns while maintaining measurement precision through statistical comparison.
3Reliability
If multiple typing samples are collected and stored, then authentication reliability is improved, but device complexity increases
Solution Approach 1:
The patent uses copying by creating a digital representation (cluster) of the user's typing patterns during enrollment. This copied reference model is then used for multiple authentication comparisons without requiring the original physical typing samples to be present, simplifying the verification process while maintaining reliability.
Solution Approach 2:
The system performs the complex task of collecting and analyzing multiple typing samples in advance during an enrollment phase, storing the results as a compressed cluster representation. This preliminary action reduces the complexity of subsequent authentication operations, which only require comparing new samples against the pre-computed cluster.
Data Source
AI summary
A system, method, and computer-readable storage device for receiving data representative of an input sample comprising a user's motion while entering information into a data input device, calculating a statistical summary of the input sample representative data, and comparing, using geometric and geospatial constructs, the calculated statistical summary of the input sample representative data to statistical summaries of representative data of previous input samples in a stored profile of a selected user to generate an indication whether the user is the selected user.