Image Organ Segmentation With Synthetic Centroid Masks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical imaging techniques, such as MRI and CT scans, struggle to accurately and efficiently isolate smaller organs like the pancreas due to their deformability, blending with surrounding tissues, and frequent overlay by other organs, making it difficult for radiologists to reliably detect their boundaries.
Innovation Solution
A heuristic-based organ segmentation method using multiple scan passes, synthetic centroid masks, and iterative thresholding to isolate smaller organs, involving frame normalization, boundary testing, and 3D correlation to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a radiologist manually views and identifies organs in medical imaging scans, then detection accuracy can be maintained through human judgment, but the process becomes time-consuming and difficult for reliably finding organ boundaries
Solution Approach 1:
The patent introduces an automated organ segmentation system that acts as an intermediary between the medical imaging scan and the radiologist. The system uses machine learning models to pre-process and segment organs in the scan, providing structured output that assists the radiologist in quickly locating and identifying organs without manual searching, thus reducing time while maintaining accuracy through automated boundary detection
Solution Approach 2:
The patent performs preliminary organ segmentation and boundary identification automatically before the radiologist views the scan. By pre-processing the medical imaging data to highlight organ locations and boundaries, the system eliminates the time-consuming manual search process while ensuring accurate detection through algorithmic boundary analysis
2Reliability
If traditional medical imaging techniques are used to detect organs like the pancreas, then the scan data can be obtained, but the homogeneity of surrounding tissue and organ position make boundary detection difficult and unreliable
Solution Approach 1:
The patent applies local quality by using region-specific analysis for organ detection. The system divides the medical imaging scan into local regions and applies specialized detection algorithms tailored to each organ's characteristics and surrounding tissue properties. This allows the system to handle the homogeneity challenge by focusing on local boundary features rather than treating the entire scan uniformly
Solution Approach 2:
The patent employs dynamic adaptive thresholding and boundary detection algorithms that adjust to local tissue characteristics. The system dynamically modifies detection parameters based on the specific organ's position, orientation, and surrounding tissue homogeneity, enabling reliable boundary detection even when organs blend with adjacent tissues
3Productivity
If smaller organs are segmented using conventional methods, then the segmentation process can be completed, but the deformability and overlay by other organs reduce detection accuracy
Solution Approach 1:
The patent uses hierarchical segmentation that divides the organ detection task into multiple stages. First, the system segments the entire organ region, then performs secondary segmentation to separate sub-structures and handle deformable regions. This multi-level segmentation approach maintains efficiency while improving accuracy for deformable organs by treating different regions with appropriate granularity
Solution Approach 2:
The patent dynamically adjusts segmentation parameters such as threshold values, smoothing factors, and boundary detection sensitivity based on organ characteristics. For deformable organs or those overlaid by other structures, the system automatically modifies parameters to enhance boundary contrast and detection accuracy, balancing segmentation speed with precise boundary identification
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Discussed herein are devices, systems, and methods for organ mask generation. A device, system and method for organ mask generation including generating a synthetic centroid mask, identifying first and second intensity thresholds, in a first segmentation pass, setting (i) pixels of an image with intensities less than the first threshold to zero and (ii) pixels of the image corresponding to objects with centroids outside the synthetic centroid mask to zero, resulting an initial organ mask, in a second segmentation pass, setting pixels (i) with intensities less than the second threshold, the second threshold less than the first threshold to zero and (ii) setting pixels corresponding to objects with centroids outside the initial organ mask to zero, resulting in a second organ mask, and expanding and filling the second organ mask to generate an organ mask.