Microwave Head Imaging Using Multi-Frequency Tomography

A microwave head imaging algorithm using multi-frequency method is proposed. The conventional microwave tomography uses only single frequency for transmission/reception to map the dielectric properties of the imaged domain. Based on the different dispersive characteristics of different brain tissues, a novel microwave tomography method using the multi-frequency approach is developed. The normalized dielectric profile of the imaging domain is firstly calculated using the Born iterative method under multi-frequency incident signals. The real and imaginary parts of the reconstructed normalized dielectric profile with different frequencies are considered as two spectrums of the imaging domain and the inverse Fourier transform is implemented on the two spectrums to obtain the time domain response for the real and imaginary parts of each tissue. The differences between the maximum values for the real and imaginary parts in the time domain are calculated and used for clustering and classification via the K-mean method. The algorithm is tested on two different numerical head phantoms and the results indicate an improved image quality compared with the traditional single-frequency tomography.