Medical Image Segmentation of Intracranial Hemorrhage: A Review

Intracranial hemorrhage (ICH) is a common clinical emergency that can lead to brain damage or death in a serious situation with extremely high disability and mortality rates. In order to obtain effective treatment for patients timely, the rapid and accurate detection of ICH is crucial. Considering that different subtypes of bleeding correspond to different treatment strategies in clinical practice, it is necessary to quickly diagnose ICH at the early stage. Computed tomography (CT) is one of the most effective ways to detect ICH. For quantifying the lesions and facilitating the subsequent treatments, the segmentation of ICH based on CT images is an extremely important step. The paper presents a review of the task of ICH segmentation, elaborates on the properties of ICH data and the available datasets for research, provides a detailed introduction to the traditional and deep-learning methods, and outlines various improvement schemes for deep-learning models. Finally, the current research achievements in this field are summarized, some urgent problems are proposed, and future research development is foreseen.