Extracting vascular networks under physiological constraints via integer programming

We introduce an integer programming-based approach to vessel network extraction that enforces global physiological constraints on the vessel structure and learn this prior from a high-resolution reference network. The method accounts for both image evidence and geometric relationships between vessels by formulating and solving an integer programming problem. Starting from an over-connected network, it is pruning vessel stumps and spurious connections by evaluating bifurcation angle and connectivity of the graph. We utilize a high-resolution micro computed tomography (μCT) dataset of a cerebrovascular corrosion cast to obtain a reference network, perform experiments on micro magnetic resonance angiography (μMRA) images of mouse brains and discuss properties of the networks obtained under different tracking and pruning approaches.