A visible human body slice segmentation method framework based on OneCut and adjacent image geometric features

Area of Science:

• Medical Imaging
• Computer Vision
• Anatomical Modeling

Background:

• Accurate 3D organ models from the Visible Human Project (VHP) are crucial for research.
• Extracting organ regions of interest (ROI) from large VHP datasets is a significant challenge.

Purpose of the Study:

• To propose a novel method framework for the continuous extraction of main organ regions from VHP image data.
• To overcome the limitations of manual delineation in processing serialized slice images.

Main Methods:

• A framework combining the OneCut algorithm with adjacent image geometric features for organ segmentation.
• Utilizing seed points from corroded foreground ROIs and skeletons from background regions as input for subsequent slice processing.
• Implementing a continuous processing pipeline for serialized slice images without manual intervention.

Main Results:

• The proposed method successfully extracts main organ regions from VHP data.
• Experimental results demonstrate satisfactory 3D organ models.
• The framework enables automated and continuous processing of slice images.

Conclusions:

• The developed method framework provides a robust technical foundation for creating accurate 3D organ models from VHP data.
• This approach facilitates efficient and automated anatomical modeling.
• The technique may support various related applications in medical imaging and research.