In the setting of an ischemic stroke, tiny blood vessels, so-called „collaterals“, are able to maintain blood flow to otherwise critically hypoperfused brain areas. These collateral vessels are of utmost importance during treatment triage of ischemic stroke patients, and knowledge about their quality and quantity are crucial for the treating physician. With this proposal, we aim to present a novel diagnostic framework to assess and quantify cerebral collateral vessels on computed tomography images using machine learning algorithms based upon artificial intelligence.