This paper proposes an efficient and fast method to create large datasets for machine learning algorithms applied to brain stroke classification via microwave imaging systems. The proposed method is based on the distorted Born approximation and linearization of the scattering operator, in order to minimize the time to generate the large datasets needed to train the machine learning algorithms. The method is then applied to a microwave imaging system, which consists of twenty-four antennas conformal to the upper part of the head, realized with a 3D anthropomorphic multi-tissue model. Each antenna acts as a transmitter and receiver, and the working frequency is 1 GHz. The data are elaborated with three machine learning algorithms: support vector machine, multilayer perceptron, and k-nearest neighbours, comparing their performance. All classifiers can identify the presence or absence of the stroke, the kind of stroke (haemorrhagic or ischemic), and its position within the brain. The trained algorithms were tested with datasets generated via full-wave simulations of the overall system, considering also slightly modified antennas and limiting the data acquisition to amplitude only. The obtained results are promising for a possible real-time brain stroke classification.
In this work, the contrast source inversion method is combined with a finite element method to solve microwave imaging problems. The paper’s major contribution is the development of a novel contrast source variable discretization that leads to simplify the algorithm implementation and, at the same time, to improve the accuracy of the discretized quantities. Moreover, the imaging problem is recreated in a synthetic environment, where the antennas, and their corresponding coaxial port, are modeled. The implemented algorithm is applied to reconstruct the tissues’ dielectric properties inside the head for brain stroke microwave imaging. The proposed implementation is compared with the standard one to evaluate the impact of the variables’ discretization on the algorithm’s accuracy. Furthermore, the paper shows the obtained performances with the proposed and the standard implementations of the contrast source inversion method in the same realistic 3D scenario. The exploited numerical example shows that the proposed discretization can reach a better focus on the stroke region in comparison with the standard one. However, the variation is within a limited range of permittivity values, which is reflected in similar averages.
The early diagnosis of Alzheimer’s disease remains an unmet medical need due to the cost and invasiveness of current methods. Early detection would ensure a higher quality of life for patients, enabling timely and suitable treatment. We investigate microwave sensing for low-cost, non-intrusive early detection and assessment of Alzheimer’s disease. This study is based on the emerging evidence that the electromagnetic properties of cerebrospinal fluid are affected by abnormal concentrations of proteins recognized as early-stage biomarkers. We design a conformal six-element antenna array placed on the upper portion of the head, operating in the 500 MHz to 6.5 GHz band. It measures scattering response due to changes in the dielectric properties of intracranial cerebrospinal fluid. A multi-layer perceptron network extracts the diagnostic information. Data classification consists of two steps: binary classification to identify the disease presence and multi-class classification to evaluate its stage. The algorithm is trained and validated through controlled experiments mimicking various pathological severities with an anthropomorphic multi-tissue head phantom. Results support the feasibility of the proposed method using only amplitude data and lay the foundation for more extensive studies on microwave sensing for early Alzheimer’s detection.
This paper presents a new and efficient method to generate a dataset for brain stroke classification. Exploiting the Born approximation, it derives scattering parameters at antennas locations in a 3-D scenario through a linear integral operator. This technique allows to create a large amount of data in a short time, if compared with the full-wave simulations or measurements. Then, the support vector machine is used to create the classifier model, based on training set data with a supervised method and to classify the test set. The dataset is composed by 9 classes, differentiated for presence, typology and position of the stroke. The algorithm is able to classify the test set with a high accuracy.
21 June 2024 in IEEE Journal of Electromagnetics, RF and Microwaves in Medicine and Biology
The contrast source inversion method is an iterative non-linear algorithm, and, in this paper, it works in combination with a finite element method solver for the reconstruction of the dielectric properties' distribution in the head with the aim to diagnose brain stroke. Here, the involved contrast source variables are discretized through a novel field-based discretization that allows a linear variation of the variables, leading to their more accurate description, and therefore to a final dielectric properties' reconstruction closer to the expected scenario. Moreover, we propose a new approach to compute the imaging algorithm initial guess, based on the truncated singular value decomposition technique, that appears more effective in the case of noisy measured data. Finally, the developed algorithm is applied to sets of data, measured with a microwave imaging system to reconstruct brain stroke scenarios.
1 January 2026 in IEEE Open Journal of Antennas and Propagation