FDTD based simulation study of a classification based hemorrhagic stroke detector

Paper in proceeding, 2018

The possibility to detect intracranial bleedings caused by stroke or head trauma in a prehospital setting would be a major breakthrough in the strive to deliver the best possible care for these patients. Our research is focused on developing a microwave based diagnostic system for prehospital use, which is capable of detecting intracranial bleedings. This paper contains a numerical simulation study to investigate the detection capability of a machine learning algorithm and its performance for differentiating hemorrhagic stroke from patients with no bleeding. The goal is to assess the performance of the detection algorithm as a function of the number of patients included in the training phase. The results show that this approach is feasible, but that it requires one thousand patients for training the algorithm, in order to reach a detection rate with AUC values approaching 0.9.