Dynamic Chest Image Waveform Analysis for Diagnosis Assistance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for pulmonary diagnosis using dynamic images struggle to efficiently acquire both physiological and morphological data, making it difficult for physicians to grasp pathological conditions, particularly under dynamic conditions, and do not adequately utilize respiratory and vascular dynamics for diagnosis.
Innovation Solution
A diagnosis assistance program that analyzes images by extracting waveforms and frequencies from dynamic images, using AI to learn and match these patterns, enabling visualization of respiratory and vascular dynamics through waveform and frequency analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional methods are used to capture dynamic chest images, then image data can be obtained, but it is difficult to efficiently acquire both physiological data and morphological data
Solution Approach 1:
The patent combines multiple diagnostic functions into a single dynamic image capture system. By simultaneously extracting morphological information (lung field regions, anatomical structures) and physiological information (respiratory waveforms, cardiovascular dynamics) from the same dynamic image sequence, the system eliminates the need for separate imaging procedures and achieves efficient multi-parameter diagnosis
2Loss of information
If maximum inter-frame difference values are displayed for each pixel, then dynamic changes are visualized, but physicians cannot easily grasp pathological conditions
Solution Approach 1:
The patent extracts specific diagnostic information from the complex dynamic image data by identifying and isolating key features: respiratory waveforms are extracted from periodic image signal variations, cardiovascular dynamics are separated from bulk motion, and morphological changes are distinguished from physiological movements. This extraction process transforms raw pixel data into clinically interpretable parameters
Solution Approach 2:
The patent applies different analysis methods to different regions of the image: lung field regions are analyzed for respiratory mechanics, pulmonary hilum regions are examined for cardiovascular dynamics, and specific anatomical structures are tracked for morphological changes. This region-specific analysis enhances the clinical interpretability of dynamic changes
3Reliability
If respiration and blood pressure are measured after the subject is placed in a resting state, then measurement can be performed, but measurement is limited to this condition and dynamic conditions cannot be assessed
Solution Approach 1:
The patent transitions from static resting-state measurement to dynamic real-time assessment. By continuously analyzing dynamic image sequences, the system captures respiratory and cardiovascular parameters during natural breathing and movement, enabling evaluation under physiologically relevant dynamic conditions while maintaining measurement reliability through real-time signal processing
4Measurement precision
If feature amounts indicating movement of divided small regions are displayed, then regional analysis is provided, but this is still not sufficient for grasping pathological conditions
Solution Approach 1:
The patent integrates multiple types of diagnostic information into a comprehensive analysis framework: morphological data from image structure, physiological data from waveform analysis, functional data from respiratory and cardiovascular dynamics, and regional data from divided small region tracking. This composite approach combines diverse information sources to provide sufficient pathological assessment
Data Source
Figure 1A
Figure 1B
Figure 1C~1D
AI summary
[Problem] To analyze a dynamic image of a living body and display an analysis result. [Solution] A diagnosis assistance program for analyzing an image of a living body and displaying an analysis result, the diagnosis assistance program causing a computer to execute a process of acquiring a plurality of frame images, a process of acquiring a periodic image signal changing at a fixed period from the plurality of frame images, a process of extracting the waveform of all or a part of the periodic image signal, and a process of collating the extracted waveform with a model waveform acquired in advance and detecting a correlation therebetween. Further included is a process of complementing discrete periodic image signals to form a continuous waveform. Further, the correlation is detected based on at least one of the phase, the amplitude, or the frequency of the waveform of all or a part of the periodic image signal.