CT Pulmonary Emboli Detection Non-Linear Contrast Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for detecting pulmonary emboli in CT images often result in false positives due to assumptions about normal anatomy, leading to incorrect identification of lymph nodes, water-filled airways, and other artifacts as emboli.
Innovation Solution
A system and method that acquire CT image data, apply a lung mask, perform non-linear contrast enhancement, filter out false positives using various techniques, and reconstruct images to isolate and output a final set of pulmonary emboli candidates, reducing false positives through a series of filtering steps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods with simple thresholds and linear contrast enhancements are used, then the processing speed is fast, but the detection accuracy deteriorates due to false positives
Solution Approach 1:
The patent transforms the linear contrast enhancement parameter into a non-linear parameter, and changes the detection threshold from a simple fixed value to an adaptive threshold based on local image statistics. This resolves the contradiction by improving detection accuracy through parameter transformation while maintaining processing efficiency through algorithmic optimization.
Solution Approach 2:
The patent performs preliminary contrast enhancement and noise filtering before the main detection process. By preparing the image data in advance with non-linear contrast enhancement, the subsequent detection algorithms work more efficiently with pre-processed data, resolving the contradiction between accuracy and processing complexity.
2Measurement precision
If large banks of information from many CT images are used to train classifiers, then the detection accuracy improves, but the processing time increases
Solution Approach 1:
The patent segments the detection process into distinct stages: contrast enhancement, candidate region identification, and verification. By dividing the processing into segments with progressively stricter filtering criteria, the system achieves high detection accuracy while minimizing the amount of data that requires intensive classification processing, thus reducing overall processing time.
Solution Approach 2:
The patent applies partial action by using simplified detection criteria for initial candidate identification, then applying more rigorous verification only to promising candidates. This avoids the need to process all image data with full classification complexity, resolving the contradiction between accuracy and processing time.
3Productivity
If simple thresholds are used for detection, then the processing is simple and fast, but false positives increase due to normal anatomy being misidentified
Solution Approach 1:
The patent applies local quality by using adaptive thresholds that vary across different regions of the image based on local tissue characteristics. Instead of a single global threshold, each region has optimized detection parameters tailored to its specific anatomical features, improving reliability while maintaining processing efficiency through localized rather than exhaustive analysis.
Data Source
AI summary
A system and method includes acquisition of a set of image data comprising a plurality of image voxels and isolation of a set of pulmonary emboli candidates from the plurality of image voxels. The system and method also includes application of a non-linear contrast enhancement to the set of pulmonary emboli candidates, filtration of the enhanced set of pulmonary emboli candidates, output of a final set of pulmonary emboli candidates, and creation of an image comprising the final set of pulmonary emboli candidates.