Inlier Neural Network for Eye Biometric Authentication Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Identity authentication systems face challenges in accurately matching interest points in eye images, leading to false positives and incorrect user identification due to inconsistent spatial positions of matched interest points.
Innovation Solution
The system employs an inlier neural network to process interest point pairs and generate inlier scores, which characterize the likelihood that interest points correspond to the same position in the eye of the same user, using a convolutional encoder neural network to generate feature maps and jointly optimize the encoder and inlier neural networks through machine learning training.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional interest point matching is used to authenticate user identity, then the authentication process can be performed, but false positives occur and authentication accuracy deteriorates due to inconsistent spatial positions of matched interest points
Solution Approach 1:
The patent introduces an inlier neural network as an intermediary component between traditional interest point matching and authentication decision-making. This neural network processes the spatial relationships of matched interest points and generates inlier scores that indicate the reliability of each match, thereby resolving the contradiction by filtering out false positives while maintaining true positives
Solution Approach 2:
The patent transforms the authentication approach by changing from direct binary matching to a probabilistic scoring system. Interest point matches are assigned continuous inlier scores ranging from 0 to 1, representing the likelihood of being a true positive. This parameter transformation enables more nuanced discrimination between genuine and false matches, improving both accuracy and reliability
2Measurement precision
If the inlier neural network is introduced to improve authentication accuracy, then false positives are reduced, but the system complexity increases
Solution Approach 1:
The patent replaces traditional mechanical interest point matching algorithms with a neural network-based system. The inlier neural network uses learned spatial relationship models to evaluate interest point pairs, substituting rule-based mechanical matching with data-driven probabilistic modeling, thereby achieving higher accuracy despite increased computational complexity
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for biometric identity authentication. In one aspect, a method comprises: determining: (i) a plurality of first interest points in a first image of an eye of a first user, and (ii) a plurality of second interest points in a second image of an eye of a second user; determining a plurality of interest point pairs; processing a network input comprising the interest point pairs using an inlier neural network to generate a network output comprising a plurality of inlier scores; and determining whether an identity of the first user matches an identity of the second user based on the inlier scores.