Adaptive Iris Matching via Partial Code Indexing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current iris recognition systems face significant challenges in processing images of poor quality, particularly in non-cooperative environments with variations in pose and lighting, leading to decreased match performance due to irregularities such as reflections, occlusions, and gazing, which existing techniques struggle to address effectively.
Innovation Solution
An adaptive iris matching approach that processes visible regions without segmenting the iris circumference, using partial encoding and database indexing to extract features and index iris information, allowing for matching without the need for upfront normalization and enabling the system to handle noisy and obscured images by adapting to local variations along radial directions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional non-adaptive matching approaches are used, then processing speed is maintained, but match performance deteriorates significantly when data quality is poor due to reflections, occlusions, and gazing
Solution Approach 1:
The system dynamically adapts its processing approach based on image quality assessment. When poor quality is detected (reflections, occlusions, gazing), the system switches from conventional non-adaptive matching to adaptive matching that processes only visible iris regions. This dynamic adaptation allows the system to maintain reliability across varying image qualities without sacrificing processing efficiency in good quality scenarios.
Solution Approach 2:
The system changes key processing parameters based on image quality: in poor quality images, it modifies the matching approach to handle partial iris visibility, adjusts normalization procedures to account for obscurations, and modifies database indexing strategies. These parameter changes enable the system to tolerate reflections, occlusions, and gazing while maintaining match performance.
2Measurement precision
If upfront normalization procedure is performed on entire iris, then matching accuracy is improved for high quality images, but processing time increases and poor quality images are rejected
Solution Approach 1:
Instead of performing normalization on the entire iris circumference, the system applies partial normalization only to visible iris regions. This partial action approach maintains matching accuracy for the observable portions of the iris while significantly reducing processing time and avoiding rejection of poor quality images where portions of the iris are obscured by reflections, occlusions, or gazing.
Solution Approach 2:
The system segments the iris processing into visible and obscured regions, applying different processing strategies to each. Visible regions undergo full normalization for accurate matching, while obscured regions are handled through adaptive matching techniques. This segmentation allows the system to maintain accuracy where possible while minimizing processing time and avoiding outright rejection of marginal quality images.
3Speed
If database indexing is performed on complete iris codes, then retrieval speed is maintained, but computational complexity increases for poor quality images with partial iris visibility
Solution Approach 1:
The system performs database indexing and search operations on partial iris codes corresponding only to visible regions rather than complete iris codes. This partial indexing approach maintains acceptable retrieval speed by operating on reduced data sets while significantly lowering computational complexity for poor quality images where only portions of the iris are visible due to reflections, occlusions, or gazing.
4Adaptability or versatility
If conventional iris recognition systems process poor quality images, then inclusivity is improved, but false match rate increases due to noise and obscurations
Solution Approach 1:
The system applies different processing qualities and strategies to different regions of the iris based on local conditions. Visible regions with good quality undergo rigorous processing with high specificity to minimize false matches, while obscured regions are processed with adaptive techniques that account for local noise and artifacts. This local quality approach allows the system to inclusively process diverse image qualities while maintaining overall reliability by not forcing uniform processing on all regions.
Data Source
AI summary
An adaptive iris matching approach for processing images of irises having a quality not sufficient for conventional non-adaptive matching approaches. Visible regions in a radial direction on an iris, without segmenting a circumferential of the iris, may be processed. A boundary of the visible region of the iris may be estimated. An iris map may be constructed with the non-visible portion of the iris masked. The iris map may be at least partially encoded. Partial codes of the iris map may be extracted to index at least portions of a database containing iris information. Irises may be retrieved from the database with an iris code as a query.