Comparison of 3-D Direct Microwave Imaging Algorithms Applied to Brain Stroke Follow-up

This paper presents features and performance comparison of two non-iterative inversion approaches used for microwave-imaging-based monitoring of after-onset stroke infarct migrations, utilizing 3-D realistic full-wave twin numerical experiments that mimic a controlled progressing hemorrhage scenario. Specifically, it considers two well-established approaches with the potential for real-time operation and low computational demand: a tomographic approach based on a truncated singular value decomposition and the Born approximation, a frequency-domain method; and the Delay-Multiply-And-Sum, a time-domain method. The setup and implications of both approaches are discussed, and their performance is assessed using various metrics, which indicate the fidelity of the retrievals.