AI-Enhanced Biomedical Antennas for 2mm Brain Tumor Detection Using Scattering, Admittance and Impedance Parameters: A Comparative Analysis

Automated detection and classification of brain abnormalities, such as tumors, is essential for understanding and tracking patient progress. In this study, advanced machine learning methods were applied to enhance the efficiency of a step-constant tapered slot antenna (STSA) for the detection and diagnosis of gliomas in the brain. The dataset used for our simulations, which were performed using the CST Microwave Studio software, consisted of tiny astrocytoma (with a tumor radius of 2 mm and a skin-tumor distance of 27 mm) and corresponding healthy controls. To simulate a human head, we employed a phantom model containing four antennas positioned orthogonally. Various machine learning algorithms were employed to identify and diagnose brain tumors. Throughout the extensive research, comprehensive analysis was conducted for the combined multimodal data, consisting of scattering (S), admittance (Y), and impedance (Z) parameters. The research findings revealed that the XGBoost algorithm, in combination with the combined S, Y, and Z parameters, achieved an impressive accuracy of 98.25%.