Automated Joint Segmentation via Thresholding and Watershed Algorithms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image segmentation methods for joint portions in medical images, particularly those using unsupervised approaches, face challenges in achieving high accuracy and reliability due to low signal-to-noise ratios and the need for manual intervention, while supervised methods rely heavily on prior information and are not adaptable to novel images.
Innovation Solution
A fully automated image segmentation method that combines the thresholding method with the watershed algorithm, utilizing a mutual supplementation relation to extract region information and generate masks, thereby segmenting joint images from skeletal medical images with high accuracy and reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual or semi-automatic segmentation methods are used, then segmentation accuracy and reliability are improved, but processing time and operational complexity increase
Solution Approach 1:
The system performs fully automatic segmentation without requiring user intervention. The algorithm automatically identifies joint regions, determines optimal thresholds, and generates segmentation masks through computational processes, eliminating the need for manual segmentation while maintaining high reliability
Solution Approach 2:
The method dynamically adjusts segmentation parameters including optimal threshold values and watershed algorithm parameters based on image characteristics. By adapting parameters to each specific image rather than using fixed values, the system achieves high accuracy while maintaining automated operation
2Productivity
If unsupervised segmentation methods are used, then processing time is reduced, but segmentation accuracy and reliability deteriorate
Solution Approach 1:
The system combines multiple segmentation approaches (thresholding method and watershed algorithm) into a unified framework. The thresholding method provides initial segmentation and the watershed algorithm refines it, merging their strengths to achieve both speed and reliability in automated segmentation
Solution Approach 2:
The method incorporates feedback mechanisms where segmentation results are evaluated and parameters are adjusted based on the output. This feedback loop ensures that the automated segmentation process continuously improves accuracy while maintaining processing efficiency
3Productivity
If thresholding method is used alone, then processing time is minimized, but region connectivity and shape information are lost
Solution Approach 1:
The method divides the segmentation process into distinct stages: initial thresholding for fast segmentation, followed by watershed algorithm application for refinement. This segmented approach maintains speed while recovering lost connectivity and shape information through the second stage
4Shape
If watershed algorithm is used alone, then morphological characteristics are captured, but excessive segmentation and loss of region information occur
Solution Approach 1:
The system applies the watershed algorithm partially rather than completely. By controlling the extent of watershed segmentation and using it only after initial thresholding, the method captures important morphological characteristics while avoiding excessive segmentation that would lose region information
Data Source
AI summary
Disclosed are a method and apparatus for fully automatically segmenting a joint based on a patient-specific optimal thresholding method and a watershed algorithm. The method of fully automatically segmenting an image may include the steps of extracting region information corresponding to a target object to be segmented from a medical image of the target object by associating a thresholding method and a load path algorithm, generating a first mask MASK 1 based on the extracted region information, generating a morphological patch by performing morphological subdivision on the medical image based on a watershed algorithm, generating a second mask MASK 2 based on the generated morphological patch, and segmenting an image corresponding to the target object from the medical image based on the first mask and the second mask.