Colon Segmentation Using Characteristic Path Cost Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for constructing a three-dimensional representation of the colon from CT scans face challenges in automatically combining disconnected fragments, particularly due to insufficient inflation or the presence of large polyps or lesions, which complicates the segmentation process.
Innovation Solution
A system and method that analyzes image scans to identify candidate segments, generates characteristic paths, connects them using a cost network, and reorients them to create a continuous volumetric virtual object of the colon, incorporating landmark segments like the rectum and cecum to ensure a well-formed path.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automatic segmentation methods are used to construct a three-dimensional representation of the colon, then processing efficiency is improved, but the accuracy of combining disconnected fragments deteriorates
Solution Approach 1:
The method segments the colon into disconnected fragments based on image intensity thresholds and anatomical landmarks, then processes each fragment independently before combining them. This allows automatic processing of multiple segments simultaneously while maintaining the ability to accurately recombine them using path-based connectivity algorithms.
Solution Approach 2:
The patent introduces characteristic paths as intermediary structures that connect disconnected colon fragments. These paths serve as mediators between fragments, allowing the system to automatically determine connectivity relationships and reconstruct the complete colon geometry without manual intervention.
2Reliability
If manual identification of relevant regions is used, then accuracy of region identification is improved, but processing time and manual intervention increase
Solution Approach 1:
The system performs self-service by automatically identifying relevant regions using image intensity thresholds and anatomical landmark detection. The algorithm independently determines which regions correspond to the colon without requiring manual annotation or radiologist intervention, thereby reducing processing time while maintaining accuracy through automated path-based verification.
3Device complexity
If disconnected fragments are present due to insufficient inflation or large polyps, then the complexity of the segmentation process increases, but the ability to reconstruct the complete colon geometry improves
Solution Approach 1:
The patent employs dynamic path adjustment algorithms that adapt to the presence of disconnected fragments caused by insufficient inflation or large polyps. The characteristic paths are dynamically recalculated based on the actual geometry of each fragment, allowing the system to handle varying levels of complexity while maintaining precise three-dimensional reconstruction of the complete colon.
Data Source
AI summary
This document discusses, among other things, systems and methods for efficiently calculating a colon segmentation from one or more candidate virtual three-dimensional objects. A sequence of image scans are analyzed and regions that represent air-filled objects and tagged-stool are identified as candidate segments. A characteristic path is generated for each candidate segment. The paths are joined using a cost network and re-oriented to be consistent with a typical flythrough path. The connected path is then used to generate a continuous volumetric virtual object.