Pulse Wave Detection Using Skin Color Segmentation and HSV-Q Signal Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing pulse wave detection technologies face challenges in robustness against disturbance elements and accuracy, particularly in non-contact, real-time monitoring scenarios such as detecting a driver's pulse wave from a moving vehicle, where facial movement and background interference complicate the detection process.
Innovation Solution
A pulse wave detection device and program that employs color space conversion, specifically using the H component for skin portion specification and the Q component for pulse wave detection, combined with real-time processing within the device, excluding disturbance elements by setting a Region of Interest (ROI) to the skin portion and correcting for brightness changes using the eye portion, enabling accurate and robust pulse wave detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a fixed evaluation region is set on the screen for pulse wave detection, then the detection method is simple, but the target moves out of the evaluation region when the target moves, making detection impossible
Solution Approach 1:
The evaluation region is transformed from a fixed static area to a dynamic tracking region that follows the target's movement. The system continuously updates the evaluation region position based on target detection, enabling real-time tracking of moving targets while maintaining simple detection operations.
2Area of stationary object
If the evaluation region includes background and non-detectable regions (hair, eyes, eyebrows, mouth), then the region coverage is comprehensive, but detection accuracy lowers
Solution Approach 1:
The system extracts only the skin portion from the comprehensive evaluation region by analyzing color space components. By identifying and isolating skin-colored pixels using H component thresholds, the method removes interfering elements such as background, hair, eyes, eyebrows, and mouth, leaving only the relevant skin area for accurate pulse wave detection.
Solution Approach 2:
Different regions within the evaluation area are treated differently based on their color characteristics. The system applies local quality filtering by identifying skin portions through color space analysis and applying detection processing only to those specific regions, rather than uniformly processing the entire evaluation region.
3Ease of manufacture
If general purpose devices are used for pulse wave detection, then the cost is low, but detection accuracy is affected by disturbance elements in practical environments
Solution Approach 1:
The system changes the parameter space by converting from RGB color space to HSV color space and utilizing the H (hue) component for skin detection. This parameter transformation enables robust distinction between skin and non-skin regions even under varying lighting conditions and environmental disturbances, maintaining high detection reliability with low-cost general purpose devices.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Improves pulse wave detection accuracy by isolating the skin portion from disturbance elements and allowing real-time monitoring, even in dynamic environments like vehicles, enhancing reliability and reducing computational load.
Implementation Method 1
it appears as a change in reflected light from the face
Implementation Method 2
taking a moving image of the face of a subject by a web camera
Data Source
Figure 1
Figure 2(a)~2(c)
Figure 3(a)~3(c)
AI summary
The purpose of the present invention is to perform pulse wave detection that is robust against disturbance factors. A pulse wave detection device 1 color converts a frame image of a moving image from RBG components to HSV components, and identifies a skin section using the skin color of a user prepared in advance with an H component. Next, the pulse wave detection device 1 converts the skin section of the frame image to YIQ components, and takes Qm as a pulse wave signal, Qm being obtained by averaging the Q values of the pixels. After carrying out the abovementioned processing on each frame image, the pulse wave detection device 1 obtains a chronological change in the pulse wave signal Qm and outputs said change as a pulse wave. As a result, it is possible to exclude disturbance factors such as the background from the pulse wave detection target, and successfully detect a pulse wave. Furthermore, the identification of the skin section is performed by using the H component to identify candidate pixels, and using the S component to narrow down the candidate pixels to the target, such that high-precision identification of the skin section is made possible.