Learning effects of automatic composition design software for human-equivalent phantoms from 1 GHz to 5 GHz with linear and exponential regression analysis

This paper describes the learning effects of automatic composition design software for human equivalent phantoms. Five materials: water, silicone emulsion, glycerin, sodium chloride, and agar were selected to compose human equivalent phantoms. Based on the adaptive linear and exponential multiple regression analysis, the developed software generates two sets of compositions for each target phantom. This paper also presents the measurement results for the eyeball, brain, and stomach phantoms designed with the developed software as examples in assessing the design accuracies and learning effects of the developed software.