Biometric Gallery Segmentation for Faster Facial Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Biometric identification systems face challenges with accuracy, latency, and throughput when dealing with large galleries, particularly with 'low-touch' biometrics like facial recognition, which are impacted by external factors and result in degraded performance as gallery size increases.
Innovation Solution
The implementation of intelligent gallery management (IGM) systems that create application-specific galleries by filtering subsets from a master enrollment gallery, optimizing gallery size for desired accuracy and latency, and using these subsets for biometric identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a large master enrollment gallery is used for biometric identification, then the system can identify a broader population, but identification time and resource usage increase significantly
Solution Approach 1:
The patent segments the large master enrollment gallery into multiple smaller application-specific galleries based on different criteria (geographic location, time period, demographic characteristics). This segmentation allows the system to process smaller subsets of data for each identification task, reducing identification time while maintaining the ability to cover the broader population through the master gallery structure.
Solution Approach 2:
The system performs preliminary actions by pre-segmenting and organizing the master enrollment gallery into application-specific galleries in advance. This pre-processing allows the system to quickly select and use the appropriate smaller gallery for a given application without having to filter through the entire large gallery during the actual identification process, thereby reducing identification time.
2Adaptability or versatility
If a large master enrollment gallery is used for biometric identification, then more people can be identified, but device resources are consumed more heavily
Solution Approach 1:
The patent divides the resource-intensive master enrollment gallery into multiple smaller application-specific galleries. Each smaller gallery requires fewer computational resources to process while collectively they maintain the population coverage capability of the original large gallery, thus reducing device resource usage for each identification operation.
Solution Approach 2:
The system applies partial action by using only the necessary subset of the master gallery (application-specific gallery) rather than processing the entire large gallery. This partial processing approach reduces resource consumption while still achieving the identification goal for the specific application context.
3Reliability
If the entire population gallery is processed for identification, then no one is missed, but accuracy degrades with larger gallery sizes
Solution Approach 1:
The patent applies local quality by creating application-specific galleries that are optimized for particular contexts (specific locations, time periods, demographics). Each local gallery has quality characteristics suited to its specific application, improving identification accuracy for that context while maintaining overall completeness through the master gallery structure.
Solution Approach 2:
The system uses feedback mechanisms to determine when to expand from application-specific galleries to the master enrollment gallery. If identification fails or confidence is low in the smaller gallery, the system can query the master gallery, using this feedback loop to maintain completeness while preserving the accuracy benefits of smaller galleries for successful identifications.
Data Source
AI summary
A system provides intelligent gallery management for biometrics. A first gallery is obtained that includes biometric and/or other information on a population of people. An application is identified. A subset of the population of people is identified based on the application. A second gallery is derived from the first gallery by pulling the information for the subset of the population of people without pulling the information for the population of people not in the subset. Biometric identification (such as facial recognition) for the application may then be performed using the second gallery rather than the first gallery. In this way, the system is improved as less time is required for biometric identification, fewer device resources are used, and so on.