Video Respiratory Abnormality Detection Using 3D Body Motion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to detect various forms of respiratory abnormalities, such as asymmetrical lung movement or abnormal chest and abdomen motion, which are not easily identified through vital signs alone and require visual observation by healthcare professionals.
Innovation Solution
An information processing system that estimates joint points from a video image, calculates relative positional relationships, and detects three-dimensional position changes of body regions to identify respiratory abnormalities, including asymmetrical lung movement and abnormal chest and abdomen motion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vital signs are used to detect respiratory abnormalities, then the detection process is simple and non-invasive, but the detection precision is insufficient for various forms of respiratory abnormalities
Solution Approach 1:
The patent replaces mechanical/vital sign-based detection with optical image analysis. By capturing images of the chest and abdomen regions and analyzing pixel brightness changes over time, the system detects respiratory movements without requiring mechanical sensors or invasive measurements, thereby improving precision while maintaining simplicity
Solution Approach 2:
The patent introduces an intermediary processing layer between the image capture and abnormality detection. By extracting feature images through correlation processing with template images, the system creates an intermediate representation that enhances the detectability of respiratory abnormalities while maintaining the simplicity of the overall system architecture
2Productivity
If visual observation by doctor or nurse is used to detect respiratory abnormalities, then the detection accuracy is high, but the productivity is low and cannot be continuously monitored
Solution Approach 1:
The system performs self-service monitoring by automatically analyzing images and detecting respiratory abnormalities without requiring human intervention. The correlation processing unit continuously compares current images with template images to detect abnormal respiratory patterns, enabling continuous monitoring while maintaining high reliability through automated decision-making
Solution Approach 2:
The patent enables continuous monitoring by processing images at regular time intervals without interruption. The system maintains continuous comparison of respiratory movements against template images, ensuring uninterrupted detection of abnormalities while improving productivity through automated continuous operation
3Measurement precision
If image processing is performed for each frame to detect respiratory abnormalities, then the detection precision is improved, but the calculation time increases
Solution Approach 1:
The patent applies preliminary action by pre-processing images to extract feature images before the main detection process. By creating template images and pre-computing correlation matrices, the system prepares processing materials in advance, reducing the computational burden during real-time abnormality detection while maintaining high precision
Solution Approach 2:
The patent extracts only the essential information needed for detection by using correlation processing to isolate feature images that contain respiratory movement information. This extraction process filters out unnecessary visual data, reducing processing time while maintaining the precision needed to detect abnormalities
Data Source
AI summary
The present technology relates to an information processing system, an information processing method, and a program that enable detection of various forms of respiratory abnormalities from a video (image) of a target person.Positions of joint points of a target person are estimated from an image of the target person captured for each frame, a relative positional relationship between a position of a body region of the target person in the image of a first frame designated by a user and the joint points estimated for the image of the first frame is calculated, the position of the body region of the target person in the image of an arbitrary frame different from the first frame is determined on the basis of the positions of the joint points estimated for the image of the arbitrary frame and the relative positional relationship, and a three-dimensional position change of the body region of the target person is detected on the basis of the position of the body region of the target person in the image of the arbitrary frame.