Neural Network Model for Tissue Thickness Estimation

This paper explores the use of optimization methods to estimate thicknesses of subcutaneous tissues using a non-invasive near-field antenna with microwave low-power reftec-tometry from 2-13.5 GHz. The initial simple model assumes a planar tissue stack and is implemented with a Long Short-Term Memory (LSTM) neural network trained on full-wave electromagnetic simulation data. A stack of tissue phantoms for skin, fat and muscle with 2 and 3 layers is used for experimental validation. The application for the work presented in this paper is non-invasive internal body temperature measurement using microwave radiometry, where knowledge of tissue thicknesses and electrical parameters is required to solve the inverse problem. These parameters vary based on body composition and type, as well as the specific placement of the sensor on the body. For example, the outermost layer of the forearm epidermis varies by 4.9 mm and upper abdominal fat can vary by 15 mm between people. The model developed for this paper predicts thicknesses of a stackup of Playdough, Rogers Duroid 6010 and a commercial skin phantom within 7 %, 42 %, and 24 % respectively, with a neural network loss value of 0.29. This preliminary model demonstrates the prediction of thicknesses of a more general N-Iayer stackup and serves as a proof of concept for a higher fidelity model that can be used to additionally predict dispersive permittivity and conductivity of layered tissues.