Driver Condition Detection Using Visible Face Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver condition detection systems face precision issues when parts of the driver's face are hidden from the camera's view, such as by the steering wheel or hands, leading to a drop in detection accuracy.
Innovation Solution
A driver condition detection system that uses a driver monitor camera to capture facial images and a detection part to focus on visible face parts like the mouth, nose, and eyes, performing matching with pre-stored facial shape data to determine the driver's condition, even when some face parts are obscured, and setting threshold values for changes in position and size to maintain accurate detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If the steering wheel turning amount increases, then the steering control capability is improved, but the face parts of the driver become hidden behind the steering wheel spokes causing detection precision to drop
Solution Approach 1:
The system segments the face detection task by identifying and selecting only the visible face parts (eyes, nose, mouth) that are not obscured by the steering wheel. The detection algorithm divides the face into multiple regions and processes only the unobscured segments to maintain detection accuracy during steering operations.
Solution Approach 2:
The system performs partial action by detecting driver condition using only the visible portion of the face rather than requiring the complete face to be visible. The detection algorithm works with incomplete facial data (partial action) by utilizing available visible features like eyes, nose, or mouth to infer driver state without needing full facial visibility.
2Ease of operation
If hands or arms obscure the face, then the driver's natural operation is maintained, but the detection precision drops due to hidden face parts
Solution Approach 1:
The detection system segments the facial region and identifies which parts are obscured by hands or arms. It then focuses processing on the unobscured segments (visible eyes, nose, or mouth) to maintain detection capability without interfering with the driver's natural hand movements or arm positions.
Solution Approach 2:
The system introduces an intermediary processing layer that analyzes the relationship between detected objects (hands, arms) and face parts. This intermediary layer determines visibility status and selectively processes only the visible portions, acting as a mediator between the obscured face and the detection algorithm.
3Measurement precision
If the system requires complete face visibility for detection, then detection accuracy is maintained, but the system becomes inoperative when face parts are hidden
Solution Approach 1:
The system performs detection using only the necessary minimum portion of the face that is visible. Rather than requiring complete face visibility, it uses partial facial features (one or more visible eyes, nose, or mouth) to perform the detection function, accepting partial action to maintain system operability.
Solution Approach 2:
The system dynamically changes the detection parameters based on visibility conditions. When the face is partially obscured, it adjusts the minimum required visible features and modifies the detection threshold parameters to maintain both accuracy and operability under varying visibility conditions.
Data Source
AI summary
The driver condition detection system includes a driver monitor camera capturing a face of a driver of a vehicle and generating a facial image of the driver, and a driver condition detection part configured to detect a condition of the driver based on the facial image. If a part of face parts of the driver is hidden in the facial image, the driver condition detection part is configured to detect a condition of the driver based on face parts of the driver not hidden in the facial image. The face parts of the driver are a mouth, nose, right eye, and left eye of the driver.