Brain Stroke Classification using a Microwave Transmission Line Approach

This paper investigates the feasibility to classify brain strokes between an intracranial haemorrhage (ICH) stroke and an ischemic stroke (IS) by using a transmission-line model. Specifically, two wideband antennas are employed to record the scattering parameters at different positions around numerical head phantoms that emulate an ICH or an IS. Then, the correlation that evaluates the similarity between the microwave signals recorded at these positions are computed and analyzed for each category of ICH and IS to support the classification. Unlike the existing methods, the proposed transmission-line approach is antenna-independent, and the classification is realized by using the measured scattering parameters directly without the need to train a classifier with extensive medical images a priori. The proposed method is studied on a few numerical head phantoms, and the results show that significant differences can be observed for the computed correlation between the ICH and IS cases.