Serial slice image segmentation of digital human based on adaptive geometric active contour tracking

Area of Science:

• Medical Imaging
• Computer Vision
• Digital Anatomy

Background:

• Manual segmentation of digital human datasets is labor-intensive and requires expert anatomical knowledge.
• Thin slice thickness in digital human data allows for contour tracking approaches.
• Existing segmentation methods may struggle with topological changes.

Purpose of the Study:

• To develop an automated and robust segmentation method for digital human data.
• To improve the efficiency and accuracy of tissue segmentation in digital human research.
• To introduce an adaptive geometric active contour tracking method.

Main Methods:

• An adaptive geometric active contour tracking method based on a feature image of object contours was developed.
• The feature image integrates contour point matching, image variance, gradient, and statistical color models.
• The method improves traditional edge-based active contour models for adaptive evolution in any direction.

Main Results:

• The proposed method demonstrates robustness in automatically handling topological changes during segmentation.
• Experimental results show the effectiveness of the method for segmenting tissues in digital human data.
• The adaptive evolution capability enhances segmentation accuracy compared to traditional methods.

Conclusions:

• The adaptive geometric active contour tracking method offers an effective solution for digital human data segmentation.
• This automated approach reduces reliance on manual segmentation and expert knowledge.
• The method shows significant potential for advancing digital human research and applications.