Compressive Sensing Techniques for Brain Stroke Monitoring

We considered the application of the compressive sensing in the microwave imaging for brain stroke monitoring. In order to detect the progress of a stroke, we used data from several measurements separated in time. The differences in the measured electromagnetic field, indicate the changes in the electromagnetic properties of the brain due to the stroke. Since these changes are small and confined to a small region, compressive sensing yields accurate images with high resolution. In this preliminary analysis, we utilized a numerical inhomogeneous head model. For simplicity, we assumed a two-dimensional (2D) geometry and the transverse-magnetic (TM) polarization. We compared the cases in which the prior knowledge of the electromagnetic parameters of the head was complete and only partial.