Eyelid Transition Time Analysis for Sleepiness Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sleepiness detection devices face challenges in accurately detecting sleepiness in individuals who intentionally blink more violently or open their eyes wider to prevent eyelid closure, as these actions can lead to increased standard deviation in eyelid opening time, making it difficult to determine sleepiness correctly.
Innovation Solution
A sleepiness detection device that computes eyelid opening/closing characteristic amounts from time series data of time intervals between eyelid transitions, using threshold values set on both the lower and upper sides of normal values to account for individual and intraindividual variations, allowing for more accurate judgment of sleepiness regardless of eyelid motion speed or vigor.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the device uses standard deviation of eye opened time to detect sleepiness, then it can detect sleepiness in most cases, but it fails to detect sleepiness when subjects intentionally blink violently or open eyes wide to prevent eyelid closure
Solution Approach 1:
The invention changes the detection parameter from standard deviation of eye opened time to eyelid opening/closing characteristic amount calculated from time series data of transition time intervals. This parameter transformation allows the system to detect sleepiness regardless of whether eyelid motion is slow or vigorous, resolving the contradiction between measurement precision and adaptability across different motion patterns
Solution Approach 2:
The invention uses dynamic analysis of eyelid transition time intervals through time series data processing. By continuously monitoring and analyzing the temporal characteristics of eyelid transitions rather than relying on static threshold values, the system adapts to various eyelid motion patterns including intentional vigorous blinking, thereby improving both detection accuracy and versatility
2Measurement precision
If the device detects four or two facial expressions to estimate awaking degree, then it can provide detailed sleepiness assessment, but it increases device complexity and cannot handle variations in facial expression correspondence to awaking degree
Solution Approach 1:
The invention extracts and focuses solely on eyelid opening/closing motion characteristics, eliminating the need to detect multiple facial expressions. By concentrating on the specific parameter of eyelid transition time intervals, the system achieves accurate sleepiness detection without the complexity of monitoring multiple facial expression types
Solution Approach 2:
The invention makes the eyelid monitoring system universal by using a single detection mechanism that can handle various sleepiness states and individual differences. The time series analysis of transition time intervals provides a unified approach that works across different subjects and sleepiness conditions, replacing the need for multiple specialized detection methods
Data Source
AI summary
There is provided a device which detects sleepiness of a human being based on the opening and closing motion of an eyelid, in which an occurrence of sleepiness of a subject is detectable with accuracy more sufficient than before even for a subject who keeps from blinking intentionally when sleepiness increases. The inventive sleepiness detecting device comprises an eyelid state detector, detecting an opened/closed state of an eyelid of a subject; a transition time interval detector, detecting sequentially a time interval between transitions between an opened state and a closed state of the eyelid; an eyelid characteristic amount computer, computing an eyelid opening/closing characteristic amount from time series data of the time intervals; a sleepiness judging device, judging that the subject feels sleepiness when the eyelid opening/closing characteristic value falls below a first threshold value or exceeds beyond a second threshold value higher than the first threshold value.