Microwave Stroke Imaging System using Learning Electric Field Enhancement

Microwave tomography is a promising method for stroke detection for its advantages of non-ionizing, portable and low cost. However, conventional microwave tomographic brain imaging requires high computation and may produce artifacts. This article proposes a Learning Electric Field Enhancement (LEFE) method for brain reconstruction and validate it using a microwave tomography imaging system. LEFE firstly employs deep learning technique to enhance the electric field, thereby increasing the degree of freedom in collected data. The reconstruction of the dielectric constant is based on the enhanced data. The proposed method reduces the reconstruction time significantly down to 0.33 seconds with an average root mean square error (RMSE) of 2.96.