Dual-Database Facial Recognition for Time-Variant Authentication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current facial recognition systems face security challenges due to human growth divergence, such as hair growth, leading to a decrease in authentication accuracy over time as the database accumulates variations, often resulting in system failure to discriminate effectively.
Innovation Solution
A dual-database facial recognition system that separates non-time-variant and time-variant data, using a baseline database for consistent facial features and a secondary database for changes like hair growth, skin tone, and accessories, with adaptive algorithms for authentication and periodic pruning of outdated data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single facial recognition database accumulates all variations over time, then the system adapts to user changes, but the security level decreases and discrimination ability is lost
Solution Approach 1:
The patent divides the facial recognition database into two separate databases: a first database storing non-time-variant enrolled facial data and a second database storing time-variant data with timestamps. This segmentation allows the system to handle different types of facial data separately, maintaining security while adapting to changes.
Solution Approach 2:
The patent implements dynamic authentication by selecting time-variant data based on current time metrics and timestamps. The system adapts its authentication process by choosing appropriate time-variant data according to when the authentication occurs, allowing flexibility while maintaining control over security parameters.
2Adaptability or versatility
If the facial recognition database accumulates many variations over time, then the system covers more user changes, but the ability to discriminate effectively decreases
Solution Approach 1:
The patent performs preliminary organization of facial data by separating time-variant and non-time-variant features during enrollment. Time-variant data is stored with timestamps, allowing the system to pre-select appropriate variations based on authentication timing, improving discrimination ability.
Solution Approach 2:
The patent changes the parameter of data selection by using time metrics to determine which time-variant data to apply. This parameter-based selection ensures that only relevant variations are considered, maintaining discrimination precision while covering user changes.
3Adaptability or versatility
If time-variant data is continuously updated and stored, then the system adapts to current user appearance, but the database complexity and processing overhead increase
Solution Approach 1:
The patent segments facial data into distinct categories (enrolled non-time-variant data and time-variant data with timestamps), simplifying the overall database structure by organizing data according to its temporal characteristics rather than storing all variations in a single complex structure.
Solution Approach 2:
The patent extracts only the necessary time-variant data features and stores them separately with timestamps, removing unnecessary complexity from the main enrolled database. This extraction approach maintains adaptability while reducing overall system complexity.
4Measurement precision
If time-variant data is used for every authentication, then authentication accuracy is maintained, but the processing time and computational resources increase
Solution Approach 1:
The patent applies partial action by selectively using time-variant data based on authentication requirements and time metrics. The system uses time-variant data only when necessary to maintain accuracy, rather than applying it universally, thus reducing processing overhead while maintaining authentication precision when needed.
Solution Approach 2:
The patent uses time metrics as a parameter to control the selection and application of time-variant data. By changing the parameter of data selection based on current time, the system optimizes the balance between authentication accuracy and processing efficiency.
Data Source
AI summary
In one aspect, a device may include at least one processor and storage accessible to the at least one processor. The storage may include instructions executable by the at least one processor to receive input from a camera indicating a first face of a first user. The instructions may also be executable to access first facial recognition data indicating one or more enrolled faces and to access second facial recognition data indicating time-variant data, where the second facial recognition data may not establish non-time-variant face data. The instructions may then be executable to select first time-variant data associated with the first user from the second facial recognition data and to authenticate the first user based on the first time-variant data and enrolled face data for the first user identified from the first facial recognition data.