Iris Normalization Using Imaginary Boundaries for Obscured Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing iris analysis methods struggle with normalization when the outer iris boundary is partly or completely obscured, such as by eyelids, which can lead to faulty identification due to pupil dilation and contraction variations.
Innovation Solution
The method involves constructing multiple imaginary outer iris boundaries based on verified estimates for the species, using these boundaries to transform digital images into normalized iris images that are insensitive to pupil variations, and comparing multiple iris codes created from these images to ensure accurate identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the outer iris boundary is determined by analyzing the digital image, then the normalization can be adapted to the actual boundary shape, but the method fails when the outer boundary is partly or completely obscured by eyelids
Solution Approach 1:
The patent applies preliminary action by constructing multiple imaginary outer iris boundaries based on verified species-specific estimates before performing normalization. Instead of waiting to detect the actual boundary (which may be obscured), the method pre-establishes multiple plausible boundary positions using known dimensional relationships between the iris and pupil for the given vertebrate species. This allows the normalization process to proceed robustly even when the true boundary is not visible in the image.
2Reliability
If multiple imaginary outer iris boundaries are constructed, then the method becomes robust to boundary obscuration, but the device complexity increases
Solution Approach 1:
The patent applies parameter changes by varying the assumed dimension of the outer iris boundary to generate multiple imaginary boundaries. Instead of complex image analysis to detect the true boundary, the method changes the boundary position parameter based on plausible dimensional estimates, creating a set of candidate boundaries that account for uncertainty and obscuration. This simplifies the overall process compared to attempting to precisely detect obscured boundaries.
3Measurement precision
If normalization is limited to certain regions of interest above and below the pupil, then faulty normalization is prevented, but the use of available iris data is reduced
Solution Approach 1:
The patent applies segmentation by dividing the iris region into multiple segments corresponding to different imaginary outer boundaries. Each segment represents a plausible extent of the visible iris data that can be used for identification. By segmenting the normalization process across multiple boundary hypotheses, the method既能 prevent faulty normalization from a single incorrect boundary assumption,又能 utilize all visible iris data across the segments for improved identification robustness.
Data Source
AI summary
The present invention pertains to a method of generating a normalized digital image of an iris of an eye for the purpose of creating an iris code for identification of vertebrates, including humans, the method comprising the steps of:capturing one or more digital images of the eye with a camera;constructing a plurality of imaginary outer iris boundaries in the one or more digital images, based on a known dimension of the outer iris boundary of the eye of a given species of vertebrates; andusing said imaginary boundaries for transforming the one or more digital images into a plurality of normalized iris image that are insensitive to variations in a dimension of a pupil of the eye.