Biometric Enrollment System Using Iris Texture Binning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Biometric identification systems face challenges in accurately enrolling and identifying subjects across a wide range of environmental conditions and over extended periods due to changes in pupil dilation, lighting, and aging, leading to increased false rejection and acceptance rates, especially in mobile applications where controlled environments are not feasible.
Innovation Solution
A biometric enrollment system that captures iris images and determines a normalized iris texture ratio (NITR), storing enrollment images in bins corresponding to different NITR ranges, allowing for adaptive matching and enrollment across varying environmental conditions, and automatically updating enrollment images based on accepted probe images to improve authentication accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a single enrollment image is used for biometric authentication, then the system is simple to operate, but authentication accuracy decreases when environmental conditions differ from enrollment conditions
Solution Approach 1:
The patent segments the enrollment process into multiple bins based on NITR ranges. Instead of using a single enrollment image, the system divides enrollment images into multiple categories (bins) according to their NITR values, allowing probe images to be matched against the most appropriate bin based on current environmental conditions, thus maintaining both simplicity and accuracy
Solution Approach 2:
The patent changes the parameter of enrollment storage by organizing images according to NITR (normalized iris texture ratio) ranges rather than storing a single image. This parameter-based segmentation allows the system to adapt to environmental changes by selecting the appropriate bin based on current lighting and pupil dilation conditions
2Reliability
If multiple enrollment images are collected under different environmental conditions, then authentication accuracy across varying conditions improves, but the enrollment process becomes more complex and time-consuming
Solution Approach 1:
The patent implements automatic bin assignment based on NITR calculation. The system automatically determines the appropriate bin for each enrollment image by calculating its NITR value and assigning it to the corresponding NITR range, eliminating the need for manual environmental control or complex enrollment procedures while still achieving accurate multi-condition authentication
3Measurement precision
If traditional normalization techniques are used to compensate for pupil dilation variations, then matching performance is maintained over small dilation ranges, but false rejection rates increase when dilation differences are large
Solution Approach 1:
The patent segments the NITR range into multiple bins, each representing a specific pupil dilation range. By dividing the continuous NITR spectrum into discrete segments, the system can select the most appropriate enrollment bin for matching, maintaining high precision across the full range of pupil dilations without the false rejections associated with traditional linear normalization
4Reliability
If the enrollment gallery size is increased to accommodate multiple environmental conditions, then authentication reliability improves, but system processing time and storage requirements increase
Solution Approach 1:
The patent segments the large enrollment gallery into multiple smaller bins based on NITR ranges. During authentication, the system first determines the probe image's NITR value and then only searches within the corresponding bin, dramatically reducing processing time compared to searching the entire gallery, while still maintaining high reliability through appropriate bin selection
Solution Approach 2:
The patent adds a dimensional organization to the enrollment gallery by introducing NITR-based binning as an additional classification dimension. This transforms the flat gallery structure into a hierarchical structure where images are first categorized by NITR range and then searched within the relevant category, improving both efficiency and reliability
Data Source
AI summary
Exemplary embodiments are directed to biometric enrollment systems including a camera and an image analysis module. The camera configured is to capture a probe image of a subject, the probe image including an iris of the subject. The image analysis module is configured to determine an iris characteristic of the iris in the probe image. The image analysis module is configured to analyze the probe image relative to a first enrollment image to determine if a match exists based on the iris characteristic. If the match exists, the image analysis module is configured to electronically store the matched probe image as an accepted image. The image analysis module is configured to select and establish the accepted image as a second enrollment image if the accepted image meets enrollment image criteria.