An End-to-End Pipeline for Vascular Network Extraction and Quantitative Characterization: Segmentation, Skeletonization, and Comparative Benchmarking

Description:

• I have been developing an integrated software framework that combines segmentation, skeletonization, and quantitative vascular analysis into a single pipeline. This framework enables automated vessel extraction, graph-based skeleton representation, and computation of key descriptors such as radius, tortuosity, curvature, volume, and surface area. By linking voxel-level image processing with graph-based modeling, my work bridges the gap between raw image data and clinically meaningful biomarkers. To support this framework, I designed custom data structures in C++ for efficient representation of vascular graphs, nodes, and edges, enabling scalable analysis of large volumetric datasets. These data structures were optimized for memory efficiency and integrated seamlessly with high-performance computing libraries such as OpenVDB and TBB. In addition, I developed visualization and interactive analysis tools using OpenGL and Dear ImGui, which allow real-time rendering, editing, and exploration of vascular networks. 3D Visualization is coming soon.....