Lung Lesion Identification via Regional Image Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for identifying lung lesions, such as those in idiopathic pulmonary fibrosis, are limited by the need for human interpretation of patterns recognized by deep learning and do not consider regions beyond the lung periphery, leading to inaccuracies in diagnosis.
Innovation Solution
An identifying device that obtains chest cross-sectional images, segments them into multiple regions from the center to the periphery, derives data for each region, and uses machine learning models to output identification results, improving the accuracy of lesion identification in the lung field.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If deep learning is used to recognize patterns of interstitial pneumonia, then pattern recognition capability is improved, but the need for human interpretation and reduced accuracy in specific disease identification occurs
Solution Approach 1:
The patent segments the lung field into multiple regions (upper lung field, lower lung field, right lung field, left lung field) and analyzes each region separately. This segmentation allows the system to capture location-specific characteristics of lesions, improving the precision of specific disease identification while maintaining the pattern recognition capability of deep learning.
Solution Approach 2:
The patent applies local quality by deriving different data characteristics for different regions of the lung field. Each region's data reflects the specific distribution and characteristics of lesions in that location, allowing the identification section to make more accurate disease identification based on location-specific patterns rather than treating the entire lung field uniformly.
2Device complexity
If only the lung periphery region is analyzed for lesion identification, then the simplicity of the analysis process is improved, but the accuracy of lesion identification deteriorates
Solution Approach 1:
The patent divides the lung field into multiple segments (upper, lower, right, left lung fields) rather than analyzing only the periphery. This segmentation enables comprehensive coverage of all lung regions while maintaining a systematic and organized analysis process, thus improving accuracy without significantly increasing complexity.
Solution Approach 2:
The patent transitions from analyzing only a single region (lung periphery) to analyzing multiple spatial regions simultaneously. This dimensional expansion from one region to four distinct regions allows the system to capture a more complete picture of lesion distribution, thereby improving identification accuracy.
Data Source
AI summary
An aspect of the present invention allows for more accurately identifying a possible lesion in a human lung field. The aspect of the present invention includes an image obtaining section configured to obtain a chest cross-sectional image of a subject, a segmentation section configured to classify, into a plurality of segments, unit elements of the chest cross-sectional image, and an image dividing section configured to divide the chest cross-sectional image into a plurality of regions. A data deriving section is configured to derive data associated with the possible lesion, the data being derived on the basis of a segment of unit elements in the each region among the plurality of segments. An identifying section is configured to output an identification result, which is a result of identification of the possible lesion in the lung field of the subject.