Microwave-based tomography of the human neck by means of a neural-network technique

In this contribution, a microwave-based tomographic approach by means of a neural network technique is applied to the human neck. The aim is to retrieve the dielectric and geometric properties of a cross section of the neck to detect the possible presence of tumors. A fullyconnected neural network and a numerical model to create a neck phantom dataset are developed. A preliminary result shows the possibility to locate and identify a simulated tumor.