Deep learning for vascular segmentation and tissue characterization in CT images

Zhang, X. (2026)

Doctoral Thesis

This thesis focuses on enhancing CTA image analysis and interpretation using the latest deep learning technologies. The overarching aim of this thesis is to overcome existing technical limitations in the automated analysis of CTA imaging data by developing advanced deep learning methods for two CTA application domains: coronary computed tomographic angiography (CCTA) and CTA of liver vasculature.

• CTA
• Coronary plaque detection
• Liver vessel segmentation
• Couinaud segmentation
• Diffusion model
• Multi-scale graph-attention
• Maximum intensity projection
• 3D graph reasoning
• Point net

• All authors: Zhang, X.
• Supervisor: Lelieveldt, B.P.F.
• Co-supervisor: Dijkstra, J.; Broersen, A.
• Committee: Quax, P.H.A.; Sánchez, C.I.; Walsum, T. van; Burgmans, M.C.
• Qualification: Doctor (dr.)
• Awarding Institution: Faculty of Medicine, Leiden University Medical Center (LUMC), Leiden University
• Date: 2026-01-07
• ISBN (print): 9789465229904
• DOI: doi:10.60602/1887-4286096