Automatic characterization and segmentation of human skin using three-dimensional optical coherence tomography

Area of Science:

• Dermatology
• Biomedical Imaging
• Computational Biology

Background:

• Accurate analysis of skin structure is crucial for dermatological research and diagnosis.
• Three-dimensional optical coherence tomography (OCT) offers detailed subsurface imaging of human skin.
• Automated analysis tools are needed to efficiently process complex OCT data.

Purpose of the Study:

• To develop and validate a set of automated algorithms for analyzing 3D OCT volumes of human skin.
• To quantify key epidermal and infundibular parameters from OCT data.
• To assess skin texture using dermal attenuation coefficients.

Main Methods:

• A suite of automated algorithms was developed for 3D OCT skin volume analysis.
• Algorithms include surface determination, epidermal segmentation, and infundibula visualization via en face shadowgrams.
• Histogram-based thresholding and distance mapping algorithms were employed for population and distribution analysis.
• Dermal attenuation coefficient calculation was performed to evaluate skin texture.

Main Results:

• The algorithms successfully determined mean epidermal thickness and generated epidermal distribution maps.
• Infundibula were visualized with high contrast, and their population and occupation ratios were calculated.
• A 3D segmented volume of infundibula was successfully generated.
• The algorithm set demonstrated high stability, portability, and reproducibility across multiple volunteer skin OCT volumes.

Conclusions:

• The developed automated algorithm set provides a robust and reproducible method for analyzing 3D OCT skin data.
• This tool enables precise quantification of skin structures, aiding in dermatological assessment.
• The algorithms show significant potential for clinical and research applications in dermatology.