Iris Recognition Viterbi Ring Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current iris recognition methods face challenges in accurately accounting for the nonlinear movement and appearance/disappearance of iris zones due to pupil dilation, leading to increased false positives and inefficiencies in identification and authentication processes.
Innovation Solution
A method utilizing a Viterbi algorithm to determine an optimal correspondence between iris rings in an acquired image and reference images, minimizing a matching distance by considering the distortion caused by pupil dilation, and accounting for the appearance or disappearance of rings through a comparison matrix and Prob functions, which can be determined through machine learning or analytical models like Clark's.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional normalization methods are used to transform iris images to fixed size, then processing efficiency is improved, but accuracy deteriorates due to nonlinear iris movement during pupil dilation
Solution Approach 1:
The iris image is divided into multiple concentric rings, and each ring is independently matched with corresponding rings from reference images. This segmentation allows the system to handle nonlinear movements and appearance/disappearance of iris zones by treating each ring separately, thereby maintaining both processing efficiency and recognition accuracy.
Solution Approach 2:
The patent implements dynamic ring matching where the correspondence between rings in the acquired image and reference images is determined optimally using the Viterbi algorithm. This dynamic approach adapts to varying pupil dilation states, allowing the system to maintain high accuracy across different imaging conditions without sacrificing processing efficiency.
2Measurement precision
If complex models like Clark's model are used to account for iris dilation, then recognition accuracy is improved, but device complexity increases due to individual-specific parameter learning
Solution Approach 1:
The system performs self-adjustment by automatically determining the optimal correspondence between iris rings using the Viterbi algorithm, without requiring pre-learning of individual-specific parameters. The Prob functions are computed on-the-fly based on the actual image data, enabling the system to achieve high accuracy while avoiding the complexity of individualized model training.
Solution Approach 2:
The patent computes Prob functions (Prob and ProbT) dynamically based on the actual ring comparisons rather than using fixed pre-trained parameters. This parameter adaptation allows the system to maintain high recognition accuracy across different individuals and conditions without requiring complex individual-specific model learning.
3Speed
If bit-by-bit comparison of iris codes is used, then processing speed is improved, but reliability deteriorates due to increased false positives from ignoring zone correlations
Solution Approach 1:
The patent implements dynamic programming using the Viterbi algorithm to determine the optimal path of ring correspondences. This dynamic approach considers the correlations between adjacent zones by evaluating multiple possible alignments simultaneously, thereby reducing false positives while maintaining efficient processing through algorithmic optimization.
Solution Approach 2:
The system maintains continuous evaluation of zone correlations throughout the comparison process by using the Viterbi algorithm to find the optimal global alignment. This continuous consideration of zone relationships, rather than isolated bit-by-bit comparison, reduces false positives while the efficient implementation keeps processing speed high.
Data Source
Figure 1~2
Figure 3a~3c
Figure 4
AI summary
The present invention relates to a method for identifying or authenticating an individual comprising the following steps implemented by a system including a storage device storing at least one binary code of a reference iris and an associated binary reference mask code: - acquisition 101 of an image of an eye of the individual, - segmentation 102 of the acquired eye image so as to isolate a region containing the texture of the iris and to determine an associated mask, - division 103 of said region containing the iris texture and of said mask into N2 concentric rings, - normalization 104 of said region containing the iris texture and of said mask from Cartesian coordinates to polar coordinates, - determination 105 of a binary iris code IC of size equal to N2*P*F and an associated binary mask code IM by applying F Gabor filters at each position among P positions of each ring,- matching 106 of at least one ring of the iris of the acquired image with a ring of the reference iris, so as to minimize a matching distance (HDViterbi) between the acquired image and the reference image, said matching distance being obtained from the determined and stored reference iris codes and the associated mask codes by calculating a sum of distances (Dist) between each of said matched rings of the acquired image iris and the corresponding ring of the reference iris, - calculation 107 of the minimized matching distance, - identification or authentication of the individual 108 if said minimized matching distance is less than a predetermined threshold.