Simulation Study of a Haemorrhagic Stroke Detector and Its Performance

Paper in proceeding, 2019

Intracranial bleedings caused by stroke or head trauma is a serious condition that need immediate medical care and interventions. Pre-hospital detection and diagnosis would constitute a major breakthrough in streamlining the care and in reducing the time from incidence to start of treatment. In this paper we present a numerical simulation study to investigate the detection capability of a machine learning algorithm and its performance when diagnosing patients with intracranial bleedings from healthy subjects, for example hemorrhagic stroke patients from healthy persons. The specific goal is to study the training phase of the classifier and how parameters, such as number of antennas, number of training samples, noise, etc. affect the ability to detect bleedings with different volumes. The detection performance is evaluated in a cross-validation scheme.