Automated Segmentation and Quality Enhancement in Medical Imaging: Applications to Cardiology

Abstract

Medical imaging is crucial for non-invasive diagnosis, treatment planning, and imageguided interventions, yet accurate analysis requires advanced processing techniques. This thesis focuses on automating segmentation in computed tomography (CT), specifically for coronary geometries and aortic calcifications. Automation enhances consistency, accelerates diagnosis, and enables scalable, reproducible analysis, facilitating data-driven and personalized clinical decision-making. By leveraging artificial intelligence, this work improves segmentation accuracy and addresses challenges such as CT artifacts. The integration of automated processes into clinical workflows optimizes operations, minimizes manual intervention, and enhances the reliability of medical imaging analysis in real-world applications.