Skin Cancer Detection Using Terahertz Metamaterial Absorber and Machine Learning

Abstract

This research aims to propose a terahertz metamaterial-based absorber that can sense the alterations in the enclosing medium’s refractive index. The suggested layout comprises a pair of concentric resonators made of gold, each resembling a ring in shape, and is mounted upon a substrate comprising of gallium arsenide (GaAs). The periodicity of the unit cell in this design is only $48~\mu $ m. At 2.47 THz, it achieves a high-quality factor (Q-factor) of 61.75 and a nearly perfect absorption of 99.50%. Parametric analyses have been performed to support the selection of the parameters used in the design. Modifications in the polarization angle do not affect the design and the absorption spectra. Because numerous biomedical samples fall within this range, the refractive index has been adjusted within the range of 1.35 to 1.39. Using the proposed sensor, 560 absorption spectra are obtained for different polarization angles, incident angles, and cell specimens of normal and cancerous skin tissue. Different machine learning algorithms have been used to classify the cells based on the absorption spectrum obtained from the proposed sensor with an accuracy of 100% and a precision and recall of 100% and 100%, respectively, on the test data. The proposed work can pave the way for future research combining machine learning and sensing at the terahertz frequency.