Shape Index Weighted Voting for Lung Nodule Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Interpreting lung CT images for detecting lung cancer nodules is challenging due to large image data volumes and distinguishing between nodules and vessels or airway walls, leading to low sensitivity and false positives in existing detection methods.
Innovation Solution
A method is introduced that calculates a response value indicative of global shape in images using local shape descriptors, spread functions, and a combination function to enhance the detection specificity of the MARS filter, which adapts to local structures and uses shape index and curvedness to differentiate between sphere-like nodules and tubular or airway structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional detection methods are used, then detection speed is maintained, but detection specificity is low and false positives are high
Solution Approach 1:
The detection process is segmented into multiple stages: initial nodule candidate detection using MARS filter, followed by shape index calculation and classification. This segmentation allows the system to apply different levels of complexity at different stages, improving overall detection specificity without overwhelming computational burden.
Solution Approach 2:
The patent applies local shape descriptors and shape index calculations specifically to regions of interest (nodule candidates) rather than the entire image. This local quality approach enhances detection precision for potential nodules while avoiding unnecessary processing of non-relevant areas, thereby improving detection specificity without proportionally increasing overall complexity.
2Measurement precision
If simple detection methods are used, then processing speed is fast, but ability to distinguish nodules from vessels or airway walls is poor
Solution Approach 1:
The MARS filter performs preliminary detection to identify nodule candidates before applying more computationally intensive shape analysis. This preliminary action filters out obvious non-nodules early, allowing detailed shape index calculations to be applied only to promising candidates, thus maintaining processing speed while improving differentiation accuracy.
Solution Approach 2:
The patent transitions from 2D image analysis to 3D volumetric analysis by calculating shape indices based on three-dimensional voxel data. This dimensional change enables better differentiation of nodules from vessels and airway walls by considering spatial relationships in all three dimensions, improving accuracy without excessively increasing processing time due to efficient 3D algorithms.
3Measurement precision
If shape analysis is added to detect global shape, then detection specificity improves, but computational complexity increases
Solution Approach 1:
Shape analysis is segmented into local shape descriptor calculation at individual voxel levels, followed by spread function application to aggregate local descriptors into global shape characteristics. This segmentation of the shape analysis process makes the complex algorithm more manageable and efficient, improving global shape detection accuracy without overwhelming computational complexity.
Solution Approach 2:
The spread function serves as an intermediary between local shape descriptors and global shape determination. It aggregates and weights local descriptor information from surrounding voxels to produce global shape indices, effectively bridging the gap between detailed local analysis and overall shape characterization while managing algorithmic complexity.
Data Source
AI summary
In one aspect of the present invention, a method for calculating a response value at a first voxel indicative of a global shape in an image is provided. The method includes the steps of (a) determining at least one local shape descriptor associated with each of the at least one local shape descriptor; (b) determining a spread function associated with the each of the at least one local shape descriptor; (c) determining second voxels around the first voxel; (d) calculating values for each the at least one local shape descriptor at each of the second voxels; (e) determining a contribution of each of the second voxels at the first voxel based on the spread functions; and (f) using a combination function to combine the contributions to determine the response value indicative of the global shape.