Driver Looking Away Detection via Dynamic Face Certainty Threshold
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing looking away detection systems fail to determine if a driver is looking away during driving when a face is not visible in the captured image, leading to incomplete detection.
Innovation Solution
A looking away determination device that uses a threshold value for face detection certainty to determine if a driver is looking away, even when a face is not visible, and updates this threshold based on the degree of certainty of face detection in images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If face detection is used to determine looking away, then detection accuracy is improved when face is visible, but detection cannot be performed when face is not visible in captured image
Solution Approach 1:
The patent introduces an imaging device that captures images of the driver as an intermediary data source. When face detection cannot determine looking away state, the system uses image processing to analyze the captured images and determine whether the driver is looking away, thus providing a fallback mechanism that expands detection coverage while maintaining accuracy through multiple detection pathways
Solution Approach 2:
The system dynamically switches between face detection method and image processing method based on detection conditions. When face detection succeeds, it uses that method for accurate determination; when face is not visible or detection fails, it transitions to analyzing captured images, creating a flexible adaptive system that maintains operational effectiveness across varying conditions
2Device complexity
If a fixed threshold value is used for face detection, then system complexity is reduced, but detection accuracy deteriorates when driving conditions vary
Solution Approach 1:
The system implements self-service through automatic threshold value updating. The updating unit automatically adjusts the threshold value based on the degree of certainty of detected faces without requiring manual intervention or complex external calibration systems. This self-adjusting mechanism maintains high detection accuracy across varying driving conditions while keeping the system relatively simple in structure
Solution Approach 2:
The system employs feedback mechanisms where the degree of certainty from face detection results is fed back to the updating unit, which then adjusts the threshold value accordingly. This closed-loop feedback system enables the threshold to adapt to changing driving conditions, lighting environments, and camera angles, maintaining optimal detection accuracy without increasing overall system complexity
Data Source
AI summary
A looking away determination device includes: a looking away determination unit that determines that a driver is in a looking away state when a proportion of an image in which a face having a degree of certainty equal to or greater than a threshold value is shown with respect to a plurality of images obtained by imaging the driver, the degree of certainty serving as an index of face likeness; and an updating unit that updates the threshold value based on a degree of certainty of a face shown in an image obtained by imaging the driver.