Eye Convergence Inattention Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional safety systems fail to detect situations where a driver is looking at the road but not paying attention to their surroundings, as they rely solely on head direction and cannot distinguish between attentive and inattentive states when the driver is daydreaming or not focusing on the environment.
Innovation Solution
An eye tracking system that uses cameras and depth sensors to determine eye convergence distance and compare it with the distance to objects in the driver's gaze, providing alerts when the convergence distance is not similar to the object distance, indicating inattentiveness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If head tracking is used to monitor driver attention, then the system can detect when the driver is looking away from the road, but it cannot detect when the driver is looking at the road but not paying attention (daydreaming or inattentive states)
Solution Approach 1:
The patent segments the attention detection task into two independent components: head position tracking and eye convergence analysis. By separating these functions, the system can first use simple head tracking to determine general direction, then apply more complex eye convergence analysis only when needed to detect inattentive states, thereby improving detection accuracy without proportionally increasing overall system complexity
Solution Approach 2:
The patent introduces eye convergence distance as an intermediary parameter that bridges the gap between simple head position monitoring and complex cognitive state assessment. By measuring where the eyes converge (using infrared cameras and pupil tracking) and comparing it to the actual object distance, the system gains a reliable indicator of visual attention without requiring direct measurement of cognitive states
2Reliability
If the system uses only head direction monitoring, then the system remains simple to implement, but it fails to distinguish between attentive and inattentive states when the driver is looking at the road
Solution Approach 1:
The patent implements a dynamic, multi-stage detection approach where the system adapts its complexity based on the situation. Head tracking provides continuous monitoring with low complexity, while eye convergence analysis is dynamically activated to provide deeper insight when head position alone is insufficient, making the system both reliable and adaptably complex
Solution Approach 2:
The patent adds a new dimension to attention monitoring by moving from two-dimensional head position tracking to three-dimensional eye convergence analysis. By measuring the convergence point of the eyes in 3D space and comparing it with object distance, the system creates a new measurement dimension that reliably indicates visual attention status without overwhelming system complexity
3Measurement precision
If eye tracking technology is implemented to detect convergence distance, then the system can identify inattentive states, but the device complexity and cost increase significantly
Solution Approach 1:
The patent makes the eye tracking system multi-functional by using the same infrared cameras and pupil tracking algorithms for both driver monitoring and potential other applications such as gesture recognition or gaze-based control. This universality justifies the added complexity by enabling multiple useful functions from the same hardware investment
Solution Approach 2:
The system uses the driver's own eye anatomy (pupil position, corneal reflection) as the tracking target, eliminating the need for external markers, contact lenses, or complex cooperative tasks from the driver. The driver's eyes essentially serve themselves as the tracking reference, simplifying the overall system despite the advanced imaging technology required
Data Source
Figure 1A~1B
Figure 2A~2B
Figure 3
AI summary
One embodiment of the present invention sets forth a technique for detecting when a user, notably the driver of a vehicle, is not being attentive based on eye convergence. The technique includes determining a first distance from the user to an obstacle, such as a preceding vehicle. The first distance may be determined using a depth sensor. The technique further includes determining a first vergence angle associated with a left eye of the user and determining a second vergence angle associated with a right eye of the user. The technique further includes determining a first eye convergence distance based on the first vergence angle, the second vergence angle, and an interocular distance between the left eye and the right eye. The technique further includes comparing the first distance to the first eye convergence distance to generate a first result. An inattention alert may be generated based on the first result. Detects if the user is looking in the direction of the object but not paying attention to it, because focussing on something else or simply daydreaming.