Real-time deep learning assisted skin layer delineation in dermal optical coherence tomography
Xuan Liu1, Nadiya Chuchvara2, Yuwei Liu1
1Department of Electrical and Computer Engineering, New Jersey Institute of Technology, University Heights, Newark, NJ 07102, USA.
OSA Continuum
|July 13, 2022
Summary
Deep learning with optical coherence tomography (OCT) enables automatic skin layer analysis. This AI-assisted imaging tool accurately measures epidermal thickness and differentiates skin lesions, aiding dermatological diagnosis.
Area of Science:
- Dermatology
- Medical Imaging
- Artificial Intelligence
Background:
- Optical Coherence Tomography (OCT) is a non-invasive imaging technique.
- Accurate skin layer characterization is crucial for dermatological diagnosis.
- Manual analysis of OCT images can be time-consuming and subjective.
Purpose of the Study:
- To develop and validate a deep learning model for automated skin layer delineation in OCT images.
- To assess the capability of AI-assisted OCT for quantitative epidermal thickness measurement.
- To evaluate the potential of this technology for differentiating normal skin from skin lesions.
Main Methods:
- Acquisition of skin OCT images using a manually scanned single fiber OCT (sfOCT) instrument.
- Training a U-Net deep learning model for automatic segmentation of skin layers.
- Quantitative analysis of epidermal thickness and comparison between normal skin and lesions.
Main Results:
- The U-Net model achieved high accuracy in automatic epidermal thickness estimation.
- AI-assisted OCT demonstrated clear differentiation between normal skin and various skin lesions.
- The sfOCT system with AI delineation showed potential for quantitative tissue assessment.
Conclusions:
- Deep learning-assisted OCT provides accurate and automated skin layer analysis.
- This technology offers a potential cost-effective tool for clinical dermatology.
- Applications include diagnosis, tumor margin detection, and objective tissue characterization.
