Iris Image Selection via Two-Stage Sharpness Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In iris recognition systems, the acquisition of usable iris images is challenging due to the lack of autofocus in embedded cameras, leading to issues with image sharpness and quality, particularly when users are not cooperative, resulting in a high rate of non-usable images being supplied to analysis and identification processing.
Innovation Solution
A method for selecting eye images that involves preselecting images through simplified sharpness and contrast analysis, followed by a detailed analysis of pupil location, intensity profile, texture sharpness, black pixel density, and pupil movement to calculate a quality index, with images exceeding a threshold being selected for further processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a simple sharpness-based selection method is used, then the system cost and processing speed are improved, but the quality of selected images deteriorates due to high rate of non-usable images
Solution Approach 1:
The patent divides the image selection process into two distinct phases: a fast pre-selection phase using simple sharpness metrics to filter out obviously poor images, and a detailed quality assessment phase applying multiple criteria (sharpness, contrast, pupil detection, iris texture analysis) only to pre-selected candidates. This segmentation allows the system to maintain high processing speed while ensuring reliable image quality for the final selection.
Solution Approach 2:
The patent implements preliminary filtering through pre-selection based on basic sharpness criteria before conducting detailed quality analysis. This preliminary action eliminates obviously unsuitable images early in the process, reducing the computational burden of subsequent detailed analysis while ensuring that only potentially usable images undergo comprehensive quality assessment.
2Reliability
If multiple detailed quality criteria are applied to all images, then the image quality selection is improved, but the computational resource consumption increases
Solution Approach 1:
The patent segments the quality assessment process into a lightweight pre-selection stage and a computationally intensive detailed analysis stage. Only images passing the pre-selection filter undergo detailed quality criteria evaluation, significantly reducing the number of images requiring complex analysis and thus lowering overall computational resource consumption while maintaining high quality standards.
Solution Approach 2:
The patent applies full detailed quality analysis only to a subset of pre-selected images rather than all captured images. This partial application of detailed criteria optimizes computational resource usage by focusing intensive analysis only where necessary, while still ensuring that selected images meet high quality standards through the multi-criteria evaluation.
3Measurement precision
If the camera is positioned close to obtain sufficient iris resolution, then the image resolution is improved, but the depth of field becomes shallow causing focus issues
Solution Approach 1:
The patent applies preliminary sharpness assessment and pre-selection filters to identify images with adequate focus quality before detailed iris analysis. This preliminary action compensates for the shallow depth of field issue by filtering out out-of-focus images captured at close distances, ensuring that only sufficiently sharp images proceed to iris recognition processing.
Solution Approach 2:
The patent replaces mechanical autofocus mechanisms with a software-based quality assessment and selection system. By using multiple quality criteria including sharpness metrics, contrast analysis, and pupil/iris detection, the system identifies and selects the best-focused images from a sequence, compensating for the lack of mechanical focus adjustment capability in embedded cameras.
Data Source
Figure 1~3B
Figure 3C~4
Figure 5~6
AI summary
The invention relates to a method for selecting images from a set of images (SV) based on sharpness and contrast criteria. The method comprises the steps of: pre-selecting images by a simplified sharpness and/or contrast analysis of each image in the set, and selecting images by a more detailed sharpness and/or contrast analysis of each pre-selected image. The invention is also applied to the selection of iris images for the purpose of iris recognition.