Thermal Face ROI Tracking for Non-Contact Vital Sign Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing non-contact image-based physiological detection methods, such as using RGB images, are not adaptable and robust due to unstable ROI caused by factors like surrounding light, face movement, and face obstructions, leading to biased detection of heart rate and respiratory rate, and thermal imaging is limited to detecting body temperature without integrating additional devices for other physiological information.
Innovation Solution
A method and system utilizing a thermal image sensor to capture dynamic physiological characteristics by applying a skeleton detection unit to locate a nose and face as an ROI, dividing it into blocks, and analyzing temperature variations to accurately detect heart rate and respiratory rate, optionally enhanced with a depth lens for 3D analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a fixed ROI is used for physiological detection, then the detection process is simple, but the detection accuracy deteriorates due to unstable ROI caused by face movement, light changes, and obstructions
Solution Approach 1:
The patent implements dynamic ROI tracking that automatically adjusts the region of interest based on real-time face position detection. The system continuously updates the ROI coordinates to follow face movements, ensuring the detection area remains accurately positioned on the face throughout the measurement period, thereby maintaining high detection accuracy despite subject movement.
Solution Approach 2:
The system employs feedback mechanisms where detection results and ROI stability are continuously monitored. When instability or inaccuracy is detected, the system automatically adjusts the ROI parameters and re-performs detection. This closed-loop feedback ensures optimal detection accuracy is maintained by adapting to changing conditions in real-time.
2Reliability
If thermal imaging is used for body temperature detection, then fever screening is effective, but other physiological information detection requires additional integrated devices
Solution Approach 1:
The patent makes the thermal imaging system multi-functional by integrating algorithms that can extract multiple physiological parameters from the same thermal images. The system processes thermal data to detect not only body temperature for fever screening but also heart rate and respiratory rate, eliminating the need for separate detection devices and reducing overall system complexity.
Solution Approach 2:
The system merges multiple physiological detection functions into a single thermal imaging platform. By combining temperature detection, heart rate monitoring, and respiratory rate analysis within one integrated system, the patent achieves comprehensive physiological monitoring without requiring multiple separate devices, thereby simplifying the overall system architecture.
3Device complexity
If RGB images are used for physiological detection, then the detection setup is simple, but the detection results are biased due to unstable ROI and environmental factors
Solution Approach 1:
The patent changes the fundamental parameter used for detection from visible light (RGB) to thermal radiation (infrared). This parameter change allows the system to detect physiological information based on temperature variations and thermal patterns, which are less affected by lighting conditions, face obstructions, and camera positioning, thereby significantly improving detection reliability while maintaining relatively simple setup requirements.
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
Accurately detects heart rate and respiratory rate without privacy invasion by capturing dynamic ROIs, improving detection precision and adaptability over traditional methods.
Implementation Method 1
applying a thermal image sensor to detect a human body and then generate a plurality of thermal images with continuous time-sequence data
Data Source
AI summary
A method for dynamic physiological characteristic region capturing includes: detecting a human body and generate thermal images with continuous time-sequence data; detecting the thermal images and locate a skeleton from one thermal images; based on the skeleton to capture a nose and a human face, the human face is set as an ROI is further divided into image blocks; relating the image blocks to the thermal images; based on variation of temperature information in continuous time-sequence data for the image blocks to divide the image blocks into the first and the second frequency variation blocks; and, analyzing the temperature information in the continuous time-sequence data to the first and the second frequency variation blocks to obtain different physiological information of the human body. In addition, a system for the same is also provided.