Dynamic Analysis System for Lung Field Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for extracting lung field regions from chest radiographic images fail to accurately capture the three-dimensional structure of lungs, leading to incomplete pulmonary function analysis and inclusion of noise, particularly missing areas behind the thoracic vertebra, heart, and diaphragm.
Innovation Solution
A dynamic analysis system that extracts an extended lung field region including both lungs and the area between them, using techniques such as contour extraction and thresholding to generate an analysis result image that accurately represents pulmonary function, thereby incorporating hidden lung regions and reducing noise influence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the lung field region is extracted using conventional thresholding methods, then the processing area is reduced and noise is excluded, but the extracted region is incomplete and misses lung areas behind the thoracic vertebra, heart, and diaphragm
Solution Approach 1:
The patent transitions from two-dimensional image thresholding to three-dimensional volumetric reconstruction. By acquiring multiple radiographic images from different angles and reconstructing the lung field in 3D space, the system captures lung regions that are obscured in 2D projections, particularly areas behind the thoracic vertebra, heart, and diaphragm. This dimensional transition resolves the contradiction by enabling complete lung field extraction without including non-lung noise.
Solution Approach 2:
The patent segments the lung field region from surrounding anatomical structures by analyzing spatial relationships and density variations in three-dimensional space. The system divides the extracted volume into lung tissue regions and non-lung regions based on their distinct radiographic properties, allowing precise separation of the lung field from the thoracic vertebra, heart, and diaphragm while maintaining completeness of the lung extraction.
2Measurement precision
If the entire image is analyzed, then all lung regions including those behind structures are captured, but noise from non-lung areas is included in the analysis
Solution Approach 1:
The patent applies different processing qualities to different regions within the extracted volume. Lung tissue regions are identified and processed with high precision using their specific radiographic characteristics, while non-lung regions are either excluded or processed with different parameters. This local differentiation allows the system to capture complete lung fields while filtering out noise from non-lung areas such as the thoracic vertebra, heart, and diaphragm.
3Productivity
If the lung field region is extracted to exclude noise, then processing efficiency is improved, but pulmonary function information from hidden lung regions is lost
Solution Approach 1:
The patent uses three-dimensional volumetric reconstruction to efficiently process complete lung fields. By working in 3D space, the system can extract the entire lung volume including hidden regions in a single processing operation, rather than attempting to extract individual 2D slices. This approach maintains processing efficiency while recovering pulmonary function information from previously inaccessible lung areas behind the thoracic vertebra, heart, and diaphragm.
Data Source
AI summary
A dynamic analysis system includes: a hardware processor that: extracts an extended lung field region formed of one region including two lungs and a region between the two lungs from a dynamic image obtained by radiography of a dynamic state of a chest of a subject; and generates an analysis result image illustrating an analysis result of a pulmonary function in the extended lung field region by analyzing the dynamic image.