Brain Stroke Classification via Machine Learning Algorithms Trained with a Linearized Scattering Operator

• DOI: 10.3390/diagnostics13010023
• Publication Date: 2022-12-22T07:06:14Z
• Authors: Valeria Mariano; Jorge A. Tobon Va...; Mario R. Casu; Francesca Vipiana
• Keywords: Perceptron; Linearization; Microwave Imaging; Operator (biology); Multilayer perceptron

This paper proposes an efficient and fast method to create large datasets for machine learning algorithms applied brain stroke classification via microwave imaging systems. The proposed is based on the distorted Born approximation linearization of scattering operator, in order minimize time generate needed train algorithms. then a system, which consists twenty-four antennas conformal upper part head, realized with 3D anthropomorphic multi-tissue model. Each antenna acts as transmitter receiver, working frequency 1 GHz. data are elaborated three algorithms: support vector machine, multilayer perceptron, k-nearest neighbours, comparing their performance. All classifiers can identify presence or absence stroke, kind (haemorrhagic ischemic), its position within brain. trained were tested generated full-wave simulations overall considering also slightly modified limiting acquisition amplitude only. obtained results promising possible real-time classification.