Driver Attention Estimation Using Road Image Region Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driving assistance systems face challenges in accurately estimating a driver's attention state, particularly in healthy states, due to the concentration of salient regions on the vanishing point in the field of view.
Innovation Solution
A driving assistance device that determines whether the vehicle is traveling in a predetermined direction, calculates a feature amount of a specific region in the photographed images where the frequency of the feature amount meeting a threshold is high, and estimates the driver's attention state based on this feature amount.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If salient region is used for estimation, then the estimation can be performed based on driver's natural line of sight, but the estimation accuracy decreases because salient regions concentrate on the vanishing point
Solution Approach 1:
The patent divides the photographed image into multiple regions (first region, second region, third region) instead of using the entire image or only the salient region. This segmentation allows the system to analyze different areas independently and combine their features, thereby improving estimation accuracy while maintaining operational simplicity.
Solution Approach 2:
The patent assigns different weights to different regions based on their importance for attention state estimation. The first region (containing the vanishing point) is given a different weight than the second and third regions. This local quality approach ensures that regions with higher diagnostic value contribute more to the final estimation, improving accuracy without complicating the overall method.
2Productivity
If only the vanishing point region is monitored, then the system focuses on the driver's natural attention area, but it cannot detect distraction states in healthy drivers
Solution Approach 1:
The patent segments the field of view into multiple regions including the vanishing point area and peripheral areas. By monitoring multiple regions simultaneously, the system can detect when a driver's attention deviates from the road (distraction) while still efficiently processing the overall scene.
Solution Approach 2:
The patent extends the monitoring from a single point (vanishing point) to multiple spatial dimensions by defining first, second, and third regions. This dimensional expansion allows the system to detect attention patterns in different areas, improving reliability for detecting distraction states in healthy drivers.
Data Source
AI summary
A driving assistance device according to the present disclosure includes a processor and a memory having instructions that, when executed by the processor, cause the processor to perform operations including: determining whether a traveling direction of a vehicle is a predetermined direction; calculating a feature amount of a predetermined region which is included in each of photographed images in the predetermined direction of the vehicle and in which a frequency at which the feature amount is equal to or larger than a threshold is equal to or higher than a predetermined frequency when the traveling direction of the vehicle is the predetermined direction; and estimating an attention state of a driver of the vehicle from the feature amount.