Interactive 3D segmentation of MRI and CT volumes using morphological operations

Area of Science:

• Medical Imaging
• Computer-Aided Diagnosis
• Image Processing

Background:

• Tomographic volume imagery analysis requires 3D rendered views for enhanced object visualization.
• The lack of effective image segmentation methods has hindered the development of 3D visualization capabilities.
• Current methods struggle to replicate complex human pattern recognition for image segmentation.

Purpose of the Study:

• To develop a novel approach for interactive image segmentation to facilitate 3D visualization of tomographic volumes.
• To circumvent the unsolved problem of automated image segmentation by integrating human user judgment.
• To demonstrate the feasibility of this interactive segmentation approach using medical imaging data.

Main Methods:

• Interactive segmentation using morphology functions guided by direct visual feedback.
• Concurrent 3D visualization during the segmentation process.
• Utilizing human user judgment and pattern recognition capabilities.

Main Results:

• Demonstrated general feasibility of the interactive segmentation approach using Magnetic Resonance (MR) and Computed Tomography (CT) images.
• The method allows users to segment images by directly interacting with morphology functions.
• Provides direct visual feedback to guide the user throughout the segmentation process.

Conclusions:

• The proposed interactive segmentation method is a feasible approach for enhancing tomographic volume imagery analysis.
• This human-in-the-loop strategy effectively bypasses the limitations of automated segmentation.
• Future advancements in computing power are expected to make this interactive 3D image processing approach practical for widespread use.