In Vivo Skin-Type Classification Using Millimeter-Wave Near-Field Probe Spectroscopy

Hecht D, Pfahler T, Ullmann I, Altstidl TR, Amer N, Jin Y, Eskofier B, Vossiek M (2022)

Publication Type: Conference contribution

Publication year: 2022

Original Authors: Damaris Hecht, Tim Pfahler, Ingrid Ullmann, Thomas Altstidl, Nadia Amer, Yi Jin, Bjorn Eskofier, Martin Vossiek

Publisher: Institute of Electrical and Electronics Engineers (IEEE)

Series: European Conference on Microwave

City/Town: Milan

Conference Proceedings Title: 2022 52nd European Microwave Conference

Event location: Milan

DOI: 10.23919/EuMC54642.2022.9924376

Abstract

This paper presents a methodology for millimeter-wave-based in vivo classification of human skin into the two naturally occurring skin types: thick skin with low water content and thin skin with high water content. For medical diagnostic applications in the microwave range, the distinction of the water content plays a major role. The method presented is based on the resonance characteristics of reflective skin measurements. Therefore, in vivo reflective spectra were measured on 10 individuals. For the measurement acquisition, a 3D printed rectangular waveguide-based near-field probe was designed that is sensitive to the skin water content in the frequency range of 75–110 GHz. In total, 360 reflection spectra were obtained, received from six different areas on the hands and arms of the volunteers. A support vector machine (SVM) model with two classes, thick skin on the palm and thin skin present on the arm and back of the hand, was trained with the measured data. The resulting SVM classifier was evaluated to have an accuracy of more than 95%. The high classification accuracy verifies the skin type classification approach as a simple and appropriate in vivo two-level hydration model for the human skin. On the basis of these results, complex challenges such as burn wound classification can become feasible, since ex vivo burn measurements have been shown to display equivalent resonant characteristics.

Authors with CRIS profile

Additional Organisation(s)

How to cite

APA:

Hecht, D., Pfahler, T., Ullmann, I., Altstidl, T.R., Amer, N., Jin, Y.,... Vossiek, M. (2022). In Vivo Skin-Type Classification Using Millimeter-Wave Near-Field Probe Spectroscopy. In Institute of Electrical and Electronics Engineers (IEEE) (Eds.), 2022 52nd European Microwave Conference. Milan, IT: Milan: Institute of Electrical and Electronics Engineers (IEEE).

MLA:

Hecht, Damaris, et al. "In Vivo Skin-Type Classification Using Millimeter-Wave Near-Field Probe Spectroscopy." Proceedings of the 2022 52nd European Microwave Conference, Milan Ed. Institute of Electrical and Electronics Engineers (IEEE), Milan: Institute of Electrical and Electronics Engineers (IEEE), 2022.

BibTeX: Download