A Novel Machine Learning Approach of Hemorrhage Stroke Detection in Differential Microwave Head Imaging System - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année :

A Novel Machine Learning Approach of Hemorrhage Stroke Detection in Differential Microwave Head Imaging System

(1) , (1) ,
1
Mohammad Ojaroudi
  • Fonction : Auteur
Stéphane Bila
M. Salimitorkamani
  • Fonction : Auteur

Résumé

In this paper, brain hemorrhage stroke detection approach using microwave-imaging system with a novel machine-learning based post-processing method is presented. In order to create a circular array based microwave imaging system sixteen elements of the modified bowtie antennas are simulated in CST medium around the full head phantom. In order to radiate in desired band from 0.5-5 GHz an appropriate matching medium is designed. In addition, a hierarchical preprocessing method is employed to calibrate the reflected signals. In the processing section, a confocal imagereconstructing algorithm based is used. Finally, a new machine learning technique including discrete wavelet transform (DWT) and principle component analysis (PCA) for feature extraction and reduction, respectively. In addition, support vector machine (SVM) is used for segmentation and clustering of hemorrhage stroke detection from reconstructed image is employed. Simulated results are presented to validate the effectiveness of the proposed method for precisely localizing and classifying bleeding targets.
Fichier principal
Vignette du fichier
1570603503.pdf (927.62 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03047739 , version 1 (08-12-2020)

Identifiants

  • HAL Id : hal-03047739 , version 1

Citer

Mohammad Ojaroudi, Stéphane Bila, M. Salimitorkamani. A Novel Machine Learning Approach of Hemorrhage Stroke Detection in Differential Microwave Head Imaging System. 2020 European Conference on Antennas and Propagation, Mar 2020, Copenhaguen, Denmark. ⟨hal-03047739⟩
122 Consultations
97 Téléchargements

Partager

Gmail Facebook Twitter LinkedIn More