A new scheme for knee joint CT image segmentation: A max-flow and watershed method driving semi-automatic CT knee...

Area of Science:

• Medical Imaging
• Computer Vision
• Orthopedics

Background:

• Knee joint segmentation from CT images is crucial for diagnosing and treating knee disorders.
• Manual segmentation is labor-intensive and requires expert knowledge, necessitating automated solutions.
• Existing automated methods may struggle with the complex anatomy and image quality of knee CT scans.

Purpose of the Study:

• To develop and evaluate a novel semi-automated scheme for 3D CT knee-joint segmentation.
• To integrate an adaptive weighted continuous max-flow algorithm with a graphical user interface (GUI) for enhanced usability.
• To improve the accuracy, efficiency, and robustness of knee joint segmentation compared to existing tools.

Main Methods:

• A semi-automated segmentation scheme combining adaptive weighted continuous max-flow and the watershed algorithm.
• Preprocessing of CT volumes for intensity normalization and robustness enhancement.
• Development of a GUI for interactive segmentation refinement and data management.
• Semi-supervised interaction allowing for limited manual input when necessary.

Main Results:

• The proposed method demonstrated superior precision, sensitivity, and specificity compared to four open-source segmentation tools across 18 datasets.
• The adaptive weighting function effectively handled weak or ill-defined edges in knee CT data.
• The GUI facilitated efficient data handling, interactive refinement, and result export.

Conclusions:

• The developed semi-automated scheme provides high-precision and efficient knee joint segmentation from 3D CT images.
• This method reduces manual effort and streamlines clinical workflows for orthopedic applications.
• The improved accuracy and robustness have the potential to enhance diagnostic capabilities and treatment planning for knee disorders.