Efficient epidermis segmentation for whole slide skin histopathological images

Area of Science:

• Digital pathology
• Medical image analysis
• Computational dermatology

Background:

• Accurate segmentation of the epidermis is essential for computer-aided diagnosis (CAD) systems in histopathological skin cancer analysis.
• Existing segmentation techniques face limitations in handling the complexity of whole slide skin images.

Purpose of the Study:

• To propose an improved computer-aided technique for segmenting the epidermis in whole slide skin histopathological images.
• To enhance the accuracy and reliability of the initial step in developing skin cancer CAD systems.

Main Methods:

• Initial segmentation using global thresholding and shape analysis.
• Application of a template matching method with adaptive template intensity.
• Final threshold calculation based on the probability density function of the processed image.

Main Results:

• The proposed technique achieved a sensitivity of 97.99% and a precision of 96.00%.
• Demonstrated superior performance compared to existing segmentation methods.
• Successfully overcame limitations of previous approaches.

Conclusions:

• The developed computer-aided epidermis segmentation technique offers improved performance for histopathological skin images.
• This advancement is critical for the development of more effective skin cancer diagnostic tools.
• The method provides a robust solution for a crucial step in digital pathology workflows.