A Compressive Sensing Application on Microwave Diffraction Tomography for the Microwave Imaging of a Stroke Affected Human Brain

Abstract

In this paper, we present a diffraction tomography based compressive sensing technique for the monitoring of hemorrhagic brain strokes. The method uses the measurements of the scattered field data on two different time instants. The difference between two measurements constitutes the data that has been used for the diffraction tomography. Since this data is naturally sparse, the compressive sensing technique has been applied to take advantage of the sparsity. The results of numerical tests with a realistic human head phantom suggest that this approach produces more accurate results with fewer antennas, compared to the conventional microwave diffraction tomography. In addition robustness of the method is examined with 30db and 40db white Gaussian noise. The simulations show that the new method is also more robust than the conventional approach. Overall, by combining compressive sensing technique with the microwave diffraction tomography, a more practical and robust technique for the differential imaging is proposed.