Thrombus localization in middle cerebral artery of patient with acute ischemic stroke on NCCT image

Acute Ischemic Stroke (AIS) is a medical condition which blood flow suddenlystop circulation to an area of the brain, resulting in brain damage. In the most case of thepatients that have AIS are in need of a treatment as soon as possible since the conditionof brain damage is worsen, and up to the point that the treatment is no longer improve thepatient’s condition, or in some cases, the treatment might decrease the patient’s condition.The result of a successful stroke treatment is a fully recover of a patient to normal with nobrain damage. On the other hand, the slow or unsuccessful treatment can result in the patientwith disability, coma, or death.A common cause of AIS is a blood clot, also known as thrombus. It obstructsthe blood flow in the middle cerebral artery (MCA). Generally, localizing the thrombus ona set of non-contrast computerized tomography (ncCT) images for some composite type isrequired qualify radiologist. In this work, we aim to automatic thrombus localization processand improve the time to find the thrombus to assist radiologist and help reduce the time foranalysis.This work presents an algorithm for localizing thrombus in MCA region on ncCTusing an image processing method. Given an ncCT image, brain region is extracted by consideringa homogeneous intensity area using k-mean cluster. A set of candidate regions is selected by considering size and average intensity of each connected region. Lastly, locationof the MCA is located to filter the candidate regions that are in the interested region. Afterthat a classification model is used to classify each candidate for either a thrombus or a braintissue