Core-Penumbra Hyperacute Ischemic Stroke Dataset

Abstract

We introduce the CPAISD: Core-Penumbra hyperAcute Ischemic Stroke Dataset, aimed at enhancing the early detection of ischemic stroke using Non-Contrast Computed Tomography (NCCT) scans. Hyperacute ischemic stroke is the earliest phase of stroke, characterized by subtle or absent findings on NCCT scans, which complicates timely diagnosis. The dataset provides a collection of segmented NCCT images, including annotations of ischemic core and penumbra regions, critical for developing machine learning models for rapid stroke identification and assessment. By offering a carefully collected and annotated dataset, we aim to facilitate the development of advanced diagnostic tools, contributing to improved patient care and outcomes in stroke management. Our dataset uniquely focuses on the hyperacute phase, where NCCT scans typically show minimal or no signs of stroke, and includes a baseline model demonstrating its utility.