Target Nodule Identification Using Seed Point Proximity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying target nodules during minimally invasive medical procedures face challenges with inaccurate segmentation due to poor image quality and irregular nodule shapes, leading to false identifications or omissions, and require either fully automated or manual approaches that are inefficient.
Innovation Solution
A hybrid system that uses image segmentation to assist in target nodule identification, allowing users to edit default shapes and boundaries suggested by the segmentation to accurately define irregular nodule shapes, and determines if a seed point is within a threshold proximity of a candidate nodule for precise identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated image segmentation is used for target nodule identification, then processing speed is improved, but identification accuracy deteriorates due to false positives from poor image quality and irregular nodule shapes
Solution Approach 1:
The system segments the nodule identification process into multiple stages: automated segmentation for initial candidate generation, user review for verification, and manual boundary adjustment for refinement. This segmentation allows automated processing to maintain speed while user involvement ensures accuracy.
Solution Approach 2:
The system introduces an intermediary user review step between automated segmentation and final identification. Users act as intermediaries who verify automated results and correct errors, bridging the gap between automated speed and manual accuracy.
2Measurement precision
If manual target nodule identification is used, then identification accuracy is improved, but processing time increases significantly
Solution Approach 1:
The system performs preliminary automated segmentation to generate candidate nodules before user review. This preliminary action handles the time-consuming segmentation task automatically, leaving users only to verify and adjust, significantly reducing total processing time while maintaining accuracy.
Solution Approach 2:
The system uses partial automated action (segmentation) combined with partial manual action (verification and boundary adjustment). This hybrid approach applies automation where it is most efficient (segmentation) while retaining manual control where it is most needed (verification), optimizing the balance between speed and accuracy.
3Device complexity
If default symmetrical shapes are used for nodule representation, then processing simplicity is improved, but representation accuracy deteriorates for irregular nodule shapes
Solution Approach 1:
The system dynamically adapts the nodule shape representation from simple default symmetrical shapes to custom fitted boundaries that match the actual irregular nodule geometry. The boundary representation evolves from static simplified forms to dynamic accurate contours based on user adjustment needs.
Solution Approach 2:
The system changes the representation parameters from fixed symmetrical geometry to variable boundary definitions. Users can adjust boundary points and shapes to accurately represent irregular nodules, changing the geometric parameters to match actual nodule morphology while maintaining processing capability.
Data Source
AI summary
A system may comprise one or more processors and memory having computer readable instructions stored thereon. The computer readable instructions, when executed by the one or more processors, may cause the system to receive image data including a segmented candidate target nodule, receive a seed point, and determine if the segmented candidate target nodule is within a threshold proximity of the seed point. Based on a determination that the segmented candidate target nodule is within the threshold proximity of the seed point, the segmented candidate target nodule may be identified as an identified target nodule. A target nodule boundary corresponding to the identified target nodule may be displayed.