Detecting hemorrhagic stroke from computed tomographic scans using machine learning models comparison

• DOI: 10.56294/dm2024.548
• Publication Date: 2025-01-16T19:14:31Z
• Authors: Zaynab Boujelb; Ahmed Idrissi; Achraf Benba; El Mahjoub Chakir
• Keywords: Stroke

Introduction: Stroke is the most leading cause of death and disability worldwide, with hemorrhagic stroke being dangerous due to bleeding in brain. To minimize impacts, early detection crucial for effective management timely intervention. This precisely motivation behind our research, which aims develop a reliable rapid diagnostic support system. Methods: In this study, authors combined machine learning (ML) models detect using dataset computerized tomography (CT) images. The study was conducted on real database containing CT images collected from Moroccan patients. method used data organization preprocessing were performed, followed by feature extraction each image, such as intensity, grayscale, histogram characteristics. These extracted features then compressed several algorithms, including Principal Component Analysis (PCA). processed fed into robust classifiers based existing literature. Results: As result, XGBoost model achieved highest classification accuracy, 93% precision, Leave-One-Subject-Out (LOSO) validation scheme. Conclusion: step forward improving patient healthcare enabling detection, could lead timely, potentially life-saving interventions.