Eyelid Detection Using Pixel Intensity Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional driver monitoring devices inaccurately detect eyelid boundaries due to makeup and shadows, leading to incorrect eye openness calculations and improper drowsiness determination.
Innovation Solution
An eyelid detection device that includes a face feature point detecting unit, an eye region detecting unit, and an upper eyelid detecting unit, which divides the eye region into areas and uses pixel intensity patterns to accurately detect the upper eyelid boundary, even in the presence of makeup or shadows.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional edge detection and curve detection are performed on captured images to detect eyelid boundaries, then the detection process is simple and fast, but detection accuracy deteriorates when makeup or shadows are present
Solution Approach 1:
The eye region is divided into multiple divided areas (first, second, third, fourth divided areas) along the upper eyelid boundary. Each divided area is processed independently to detect boundaries between upper eyelid portions and eye portions, then these boundaries are connected to form the complete upper eyelid line, improving detection accuracy while maintaining manageable complexity
Solution Approach 2:
Different detection methods are applied to different divided areas based on their local characteristics. The first and second divided areas use one detection approach, while the third and fourth divided areas use another approach, allowing the system to adapt to local variations in makeup and shadow patterns
2Measurement precision
If the eye region is divided into multiple areas for detailed analysis, then detection accuracy improves, but processing time and computational load increase
Solution Approach 1:
The eye region is segmented into four specific divided areas arranged in the vertical direction, allowing parallel processing of multiple regions. This segmentation enables the system to process different areas simultaneously rather than sequentially, reducing overall processing time while maintaining high detection accuracy through localized analysis
Solution Approach 2:
The boundaries detected in each divided area are merged and connected to form the complete upper eyelid line. This combining step consolidates the results from multiple divided areas into a single coherent detection result, achieving accurate boundary detection without requiring excessively complex individual area processing
Data Source
AI summary
An eyelid detection device includes a face feature point detecting unit for detecting outer and inner eye corners as face feature points from a captured image, an eye region detecting unit for detecting an eye region including both an upper eyelid area and an eye area from the captured image on the basis of the face feature points, and an upper eyelid detecting unit for, for each of a plurality of divided areas into which the eye region is divided, detecting the boundary between the upper eyelid area and the eye area on the basis of the pattern of arrangement of the intensity values of pixels aligned in the divided region, and for connecting the boundary detected for each of the divided regions, to determine an upper eyelid.