Lung Field Area Division for Diffuse Disease Case Search

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The lung field area's complex and asymmetric shape makes it difficult to accurately divide into suitable areas for image diagnosis, particularly for diffuse lung disease, where lesions are distributed over a wide range, leading to challenges in searching for similar cases.

Innovation Solution

A method that extracts the lung field area, identifies its contour including the chest wall and mediastinum, internally divides it into central and peripheral areas based on shape, counts lesion pixels in each area, and refers to a storage unit for similarity levels to find matching cases.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If the lung field area is divided based on distance from the extracted center, then the division process is simple, but the division accuracy is insufficient due to the complicated and asymmetric shape of the lung field area

Engineering Contradiction:
Improvedivision process simplicityVSAvoiddivision accuracy
Core Design Contradiction:
Ease of manufactureVSManufacturing precision

Solution Approach 1:

The lung field area is segmented into multiple regions (central area, peripheral area, and intermediate area) based on the contour shape rather than simple radial distance. The contour is divided into a plurality of regions, and pixels are allocated to different areas based on which contour region they fall within, providing both structural clarity and diagnostic utility.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different regions of the lung field are assigned different functional qualities based on their anatomical and diagnostic characteristics. The central area, peripheral area, and intermediate area each have distinct diagnostic significance for diffuse lung diseases, allowing radiologists to evaluate lesion distribution patterns in functionally meaningful zones.

Inventive Principle:
Principle #3Local quality

2Measurement precision

If the lung field area is divided into suitable areas for image diagnosis, then the accuracy of lesion distribution identification is improved, but the device complexity increases

Engineering Contradiction:
Improvelesion distribution identification accuracyVSAvoiddivision method complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The contour of the lung field area is extracted and pre-divided into multiple regions before lesion analysis. By establishing the regional framework in advance based on contour geometry, the system prepares the anatomical map that will guide subsequent lesion pixel classification, improving efficiency and accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system creates a simplified representational model of the lung field by copying its contour geometry and dividing it into standardized regions. This geometric model serves as a template for classifying lesion pixels without requiring complex anatomical knowledge or manual intervention during the analysis phase.

Inventive Principle:
Principle #26Copying

3Productivity

If the lung field area is not accurately divided, then the search for similar cases becomes faster, but the reliability of similar case search results deteriorates

Engineering Contradiction:
Improvesearch speedVSAvoidsimilar case search accuracy
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system changes the parameter basis for division from simple radial distance to contour-based regional classification. By using the actual contour geometry to define areas, the system captures the true anatomical variation in lung field shape, improving the accuracy of lesion distribution characterization while maintaining automated processing speed.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3594830B1Similar case image search program, similar case image search device, and similar case image search method
Publication Date: 2021.05.26 FUJITSU LTD
  • EP3594830B1 patent drawingFigure 1
  • EP3594830B1 patent drawingFigure 2
  • EP3594830B1 patent drawingFigure 3A

AI summary

A similar case image search method performed by a computer, the method includes: extracting a lung field area from a medical image and identifying a contour of the lung field area including a chest wall and a mediastinum; identifying a position at which the chest wall and the mediastinum are internally divided and dividing the lung field area into a central area and a peripheral area based on a shape of the lung field area; counting the number of pixels indicating lesions in each of the divided central area and peripheral area; and identifying a similar case image corresponding to similarity level of the number of pixels indicating lesions by referring to a storage unit that stores the number of pixels indicating lesions in each of the areas.