Eye State Detection Using Iris Shape and Eyelid Curve Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current eye open or closed state detection methods have low accuracy due to reliance on a single threshold for eye opening measurement, which is unstable and affected by environmental factors and variations in eye opening distributions among individuals.
Innovation Solution
The method involves obtaining a grayscale image of an eye, extracting eye features including iris shape and vertical direction features, and using a classification model to determine the eye open or closed state, improving detection accuracy by considering multiple features rather than just eye opening.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single threshold is used to determine eye open or closed state based on eye opening measurement, then the detection process is simple, but the detection accuracy is low due to instability and environmental factors
Solution Approach 1:
The patent transitions from a single-dimensional eye opening measurement to a multi-dimensional feature space by introducing iris shape features (circularity, aspect ratio, area) and vertical direction features (eyelid curve changes). This dimensional expansion allows the system to differentiate eye states more accurately by considering multiple characteristics simultaneously rather than relying on a single threshold value.
Solution Approach 2:
The patent changes the detection parameters from a single eye opening measurement to multiple parameters including iris shape parameters (circularity, aspect ratio, area) and vertical direction parameters (eyelid curve derivatives). This parameter transformation enables more robust detection by capturing different aspects of eye morphology that remain stable across environmental variations and individual differences.
2Measurement precision
If multiple features including iris shape and vertical direction features are extracted and used for classification, then detection accuracy is improved, but the feature extraction and processing complexity increases
Solution Approach 1:
The patent segments the eye detection task into distinct feature extraction components: eye opening feature extraction, iris shape feature extraction (including circularity, aspect ratio, area calculations), and vertical direction feature extraction (eyelid curve analysis). This segmentation allows each feature to be computed independently using dedicated algorithms, making the overall complex process more manageable and enabling parallel processing optimization.
Solution Approach 2:
The patent develops a multi-functional classification model that processes multiple types of features (eye opening, iris shape, vertical direction) through a unified SVM framework. This universal classifier handles diverse feature inputs and produces a single eye state determination, reducing the need for separate processing pipelines for each feature type and simplifying the overall system architecture despite the increased feature dimensionality.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
This application provides an eye open or closed state detection method and an electronic device, to resolve an existing problem of relatively low accuracy of determining an eye open or closed state merely by comparing an eye opening with a single threshold. In the embodiments of this application, in addition to an eye opening feature, an iris shape feature and/or a vertical direction feature are added to identify the eye open or closed state, where the eye opening feature is used to represent an eye opening degree, the iris shape feature is used to represent a shape of an iris of an eye, and the vertical direction feature is used to represent a change degree of an eyelid curve. The eye open or closed state is identified by using a plurality of features, thereby improving detection accuracy.