Flash Eye Defect Detection Using Dynamic Anthropometric Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digital image processing algorithms are inadequate in effectively detecting and correcting flash-induced eye defects in digital photographs, particularly due to their rudimentary nature and failure to account for physiological variations and capture conditions, leading to both false positives and missed true positives.
Innovation Solution
The use of image acquisition data, such as flash position, distance, focal length, ambient light, and dynamic anthropometric data, to determine the appropriate defect correction algorithms and parameters for each image, allowing for more precise identification and correction of flash-induced eye defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If rudimentary algorithms are used to detect flash-induced eye defects, then the device complexity is reduced, but the measurement precision and reliability of defect detection deteriorate
Solution Approach 1:
The patent changes multiple parameters simultaneously: uses RGB color space with specific wavelength ranges (600-700nm for red, 450-500nm for blue), applies specific color ratio thresholds (R/B ratio > 1.5), incorporates geometric parameters (circularity, aspect ratio), and uses physiological parameters (pupil size 2-8mm, iris diameter 3-10mm). This multi-parameter approach significantly improves detection precision while maintaining reasonable computational complexity.
Solution Approach 2:
The patent transitions from simple 2D color-based detection to a multi-dimensional detection framework that includes: color space (RGB wavelengths), geometric dimensions (circularity, aspect ratio, area ratios), physiological dimensions (pupil-to-iris ratio, eye-gaze angle), and spatial dimensions (flash-to-lens distance, subject distance). This dimensional expansion enables precise differentiation between true defects and false positives.
2Object-affected harmful factors
If narrow color spectrum and shape range filters are applied to identify defect candidates, then false positives are reduced, but true positives are also missed
Solution Approach 1:
The patent employs a multi-parameter threshold system that evaluates multiple conditions simultaneously: color ratios (R/B > 1.5, G/B < 0.8), geometric parameters (circularity 0.7-1.0, aspect ratio 0.8-1.2), and physiological parameters (pupil-to-iris ratio 0.3-0.7). This multi-constraint approach maintains high true positive detection while filtering false positives, as false positives rarely satisfy all parameters simultaneously.
Solution Approach 2:
The patent introduces dynamic adaptation mechanisms where detection thresholds are adjusted based on image-specific conditions such as lighting conditions, subject distance, and flash-to-lens distance. The system dynamically determines acceptable parameter ranges rather than using fixed thresholds, enabling reliable detection across varying conditions without increasing false positives.
3Device complexity
If algorithms do not account for physiological variations and capture conditions, then the algorithm simplicity is maintained, but the adaptability to different conditions deteriorates
Solution Approach 1:
The patent incorporates physiological parameters (pupil diameter 2-8mm, iris diameter 3-10mm, pupil-to-iris ratio 0.3-0.7, eye-gaze angle -30 to +30 degrees) and capture condition parameters (flash-to-lens distance, subject distance, ambient light levels) into the detection algorithm. These parameters are processed using standardized calculations that maintain algorithmic simplicity while significantly improving adaptability to different physiological variations and capture conditions.
4Measurement precision
If comprehensive image processing with multiple parameters is performed, then the detection precision is improved, but the processing time and device complexity increase
Solution Approach 1:
The patent segments the detection process into distinct independent modules: color analysis module (calculating RGB ratios), geometric analysis module (computing circularity and aspect ratios), physiological analysis module (determining pupil-to-iris ratios and eye-gaze angles), and synthesis module (combining all parameters for final determination). This segmentation allows parallel processing of independent parameters, reducing overall processing time while maintaining comprehensive analysis for high detection precision.
Data Source
Figure 1a~1d
Figure 2
Figure 3a~3b
AI summary
A method and device for detecting a potential defect in an image comprises acquiring a digital image at a time; storing image acquisition data, wherein the image acquisition data includes at least one of a position of a source of light relative to a lens, a distance from the source of light to the lens, a focal length of the lens, a distance from a point on a digital image acquisition device to a subject, an amount of ambient light, or flash intensity; determining dynamic anthropometric data, wherein the dynamic anthropometric data includes one or more dynamically changing human body measurements, of one or more humans represented in the image, captured at the time; and determining a course of corrective action based, at least in part, on the image acquisition data and the dynamic anthropometric data.