Brain Stroke Microwave Imaging via an Efficient Implementation of the CSI-FEM Algorithm

Microwave imaging of the human head for stroke detection is demonstrated using the finite-element contrast source inversion method with enhanced discretization of the contrastsource variable. The linear basis functions used in the new discretization lead to a simple implementation of higher accuracy compared to discretizations wherein the contrast source variable is assumed to be constant over each tetrahedron of the 3D finiteelement mesh. These advantages are particularly important for stroke imaging because of the highly inhomogeneous nature of the human head. Results using synthetic data obtained from a realistic numerical model of the head show promise for stroke detection.