Machine Learning Classification of Human Osseous Tissue through Microwave Sensing

—Globally, microwave frequencies are being extensively employed in numerous biomedical implementations due to its high resolution, reasonable penetration through the human tissue, and cost-effectiveness. However, the quantization of human osseous tissue through microwave sensing is still not proficient. Therefore, this article provides an insight on the prediction of onset and progression of osteoporosis developed through the use of a microwave setup for the contactless evaluation of osteoporosis. This microwave setup comprises a human wrist model as a device under test which is illuminated through a pair of planar stubbed monopole antennas to characterize the different degrees of osteoporosis through frequency domain simulation analysis. By diversifying the wrist dimensions, we are collecting the dataset of the transfer characteristics. Furthermore, different machine learning algorithms are employed on this dataset to train, classify, and eventually evaluate the different degrees of osteoporosis. Finally, an optimum machine learning algorithm was obtained to work at an optimum bandwidth and optimum frequency. © 2023, Electromagnetics Academy. All rights reserved.