Hexagonal Tiled Terahertz Metamaterial Absorber for Cancer Detection Incorporating Machine Learning

This research paper presents the design of a hexagonal tiled metamaterial absorber for the detection of blood and breast cancer incorporating machine learning techniques. A simple hexagonal tiled metamaterial absorber with 5 × 5 concentric hexagonal rings in a unit cell has been designed for sensing the refractive index ( n ) of blood cancer cells ( n =1.390), normal blood cells ( n =1.376), breast cancer cells ( n =1.399) and normal breast cells ( n =1.385) in the terahertz region. The absorber offers four distinct peaks in its absorption rate, i.e., 78% at 0.7542 THz, 99.41% at 2.2333 THz, 99.57% at 3.2901 THz, and 94.04% at 3.2915 THz, respectively, at unity refractive index and three distinct peaks in the refractive index range of 1.3 - 1.4. The sensitivity and Figure of Merit of two peaks with maximum absorption rate were found to be 0.427 THz/RIU and 5.035, 0.6847 THz/RIU, and 4.383, respectively, where RIU stands for Refractive Index Unit. Quality Factor was calculated to be 17, 24, and 23, respectively. The Full Width Half Maximum (FWHM) bandwidth was determined to be 0.0419, 0.0848, and 0.1562 THz, respectively. Moreover, a huge data set was generated by varying the thickness of the analyte and refractive index to detect cancer cells using machine learning. With extra tree classifier the detection accuracy was found to be 99.62% on train data and 99.52% on validation data. Thus, this proposed absorber can be efficiently used to detect blood and breast cancer.