Pulmonary Nodule Image Extraction Using Adaptive Decision Boundary
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing methods struggle to accurately extract regions of bronchi and arteries/veins around pulmonary nodules in medical images due to their thin diameter and low contrast with the surrounding lung field, and the movement caused by pulsation, leading to blurred and broken images.
Innovation Solution
An image processing apparatus that extracts pulmonary nodules and sets a decision boundary at a distance determined by the nodule's size, then identifies and acquires regions of bronchi or arteries/veins that contact both the nodule and the boundary, ensuring accurate extraction by considering the anatomical structure and direction of these features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image processing methods are used to extract bronchi and arteries/veins, then all structures in the lung field can be extracted, but the extraction accuracy is low due to thin diameter and low contrast
Solution Approach 1:
The patent segments the extraction process into multiple stages: first extracting the pulmonary nodule region, then using it as a reference to extract surrounding bronchi and arteries/veins. This segmented approach focuses computational resources on the specific area of interest, improving extraction accuracy for thin structures with low contrast.
Solution Approach 2:
The patent applies different extraction strategies locally: using the pulmonary nodule as a reference point to guide the extraction of adjacent structures. The decision boundary is set at a specific distance from the nodule, creating a localized extraction zone that prioritizes accuracy for structures in proximity to the nodule while maintaining computational efficiency.
2Productivity
If conventional methods extract all bronchi in the lung field, then comprehensive coverage is achieved, but processing time and computational load increase
Solution Approach 1:
The patent extracts only the relevant portion of the lung field by using the pulmonary nodule as a reference. Instead of processing the entire lung field, the method extracts bronchi and arteries/veins within a specific distance from the nodule, significantly reducing processing time and computational load while maintaining clinical relevance.
Solution Approach 2:
The patent performs preliminary extraction of the pulmonary nodule region before extracting the surrounding structures. This preliminary action establishes a reference frame and decision boundary that guides subsequent extraction, avoiding the need to process the entire lung field and improving overall processing efficiency.
3Measurement precision
If the decision boundary is set close to the pulmonary nodule, then extraction precision is improved, but false positives from pulsation-induced blurring increase
Solution Approach 1:
The patent introduces a decision boundary as an intermediary element between the pulmonary nodule and the extraction region. This boundary is set at an optimized distance that balances precision and reliability, acting as a filter that reduces false positives from pulsation-induced blurring while maintaining accurate extraction of relevant structures.
4Reliability
If extraction distance from pulmonary nodule is increased, then false positives from pulsation are reduced, but extraction precision decreases
Solution Approach 1:
The patent optimizes the extraction distance parameter based on the pulmonary nodule size and characteristics. By dynamically adjusting this parameter rather than using a fixed distance, the method achieves an optimal balance between reducing false positives from pulsation and maintaining extraction precision for clinically relevant structures.
Data Source
AI summary
A first extraction unit that extracts a region of a pulmonary nodule drawn in an image, a second extraction unit that extracts a region of a structure drawn in the image, a setting unit that sets a decision boundary to be at a position away from the pulmonary nodule region by a distance determined on a basis of a size of the pulmonary nodule region, a decision unit that decides whether or not the structure region extracted by the second extraction unit contacts with both a region related to the pulmonary nodule region and the decision boundary, and an acquisition unit that acquires, as a region of a desired structure, the structure region decided by the decision unit to contact with both the pulmonary nodule region and the decision boundary in the structure region extracted by the second extraction unit are included.