Dynamic Identity Authentication With Non-Local Motion Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Legacy authentication technologies face challenges in providing ease of use and quality of authentication, particularly in complex modern scenarios, as seen with the implementation delays of Strong Customer Authentication (SCA) specifications under the revised European Payment Services Directive (PSD2) for open banking initiatives.
Innovation Solution
A dynamic identification method (DYNAMIDE) that utilizes spatiotemporal trajectories of anatomical landmarks during activities, processed by non-local graph convolution neural networks, to identify individuals based on their unique motion patterns, reducing the need for traditional biometric features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If legacy multi-factor authentication procedures are used, then authentication security is maintained through multiple factors, but ease of use deteriorates due to complex procedures and multiple challenges
Solution Approach 1:
The patent extracts and removes the need for multiple authentication factors by identifying and utilizing unique idiosyncratic motion patterns inherent in human anatomy. Instead of requiring multiple separate authentication factors, the system extracts a single biometric identifier from natural, unconscious motion patterns during routine activities, thereby maintaining security while eliminating procedural complexity
Solution Approach 2:
The system enables self-service authentication by capturing unconscious, natural motion patterns that individuals perform automatically during daily activities. The authentication process requires no active participation or conscious effort from the user beyond performing their normal routine actions, as the system passively observes and analyzes idiosyncratic motion characteristics
2Measurement precision
If traditional biometric features are used for authentication, then identification accuracy is achieved, but adaptability deteriorates due to rigid authentication requirements
Solution Approach 1:
The patent transitions from static biometric features to dynamic motion pattern analysis. By capturing and analyzing spatiotemporal trajectories of anatomical landmarks during movement, the system creates a living, dynamic biometric identifier that adapts to natural variations in human motion while maintaining high identification accuracy through consistent idiosyncratic patterns
Solution Approach 2:
The system changes the fundamental parameters of biometric authentication from static physical measurements to dynamic spatiotemporal motion characteristics. This parameter transformation enables the system to accommodate natural variations in human behavior and anatomy while maintaining precise identification through unique motion signatures
3Productivity
If apriori processing constraints are applied to AFID trajectories, then processing efficiency is improved, but measurement precision deteriorates due to limited trajectory features
Solution Approach 1:
The patent applies partial processing constraints that focus computational resources on the most discriminative aspects of motion patterns. Rather than processing all possible trajectory features equally, the system selectively analyzes specific spatiotemporal characteristics that provide the highest identification value, achieving both efficiency and precision
Data Source
AI summary
A method of identifying a person, the method comprising: acquiring spatiotemporal data for each of a plurality of anatomical landmarks associated with an activity engaged in by a person that defines a spatiotemporal trajectory of the anatomical landmark during the activity; modeling the acquired spatiotemporal data as a spatiotemporal graph (ST-Graph); and processing the ST-Graph using at least one non-local graph convolution neural network (NLGCN) to provide an identity for the person.