Ghana Brain CT Stroke Development Dataset (V1): 5,744 Non-Contrast Computed Tomography Images from 284 Patients

The Ghana Brain CT Stroke Development Dataset (V1) consists of 5,744 non-contrast computed tomography (NCCT) brain images obtained from 284 patients of Spectra Health, Kumasi, Ghana. The aim of developing this dataset is to support the progress in the fields of artificial intelligence, machine learning, computer vision, and image analysis for automatic stroke diagnosis and classification. All images were acquired in the course of regular medical treatment and then anonymized to preserve the privacy of patients. The dataset includes images from No Stroke, Ischemic Stroke, and Hemorrhagic Stroke studies, making it a useful tool for designing computer-assisted diagnostic algorithms. The dataset will be especially helpful in studying deep learning, XAI, transfer learning, image classification, segmentation, and medical decision making. This dataset is one of a few publicly available brain CT datasets that come from Sub-Saharan Africa, thus adding variety to existing medical imaging data. Moreover, it can serve as a development version of a larger Ghanaian stroke imaging dataset to be used in further researches. File names with n(number) means no stroke, those with s(number) have stroke