Benchmark Diagnostic MRI and Medical Imaging Dataset

This medical image dataset provides a well-off collection of images to enhance the training of computer vision and artificial intelligence models for accurate diagnosis. By supporting the early detection of diseases, it accelerates research into common and rare conditions, fostering advancements in medical imaging technologies.

We collected 34,192 medical images with 40 classes of diseases focused on MRI scans of the brain, neuro, and spine, sourced from multiple hospitals. The dataset includes 40 classes, with 26 MRI brain classes and additional classes covering neuro, spinal cord, and systemic disorders. The dataset includes several rare diseases, such as Walker-Warburg Syndrome (808), Pachygyria with Cerebellar Hypoplasia (592), Moyamoya Disease with Intraventricular Hemorrhage (864), Hallervorden-Spatz Disease (Pantothenate Kinase-Associated Neurodegeneration) (1064), and Fukuyama Muscular Dystrophy (920). It also includes more common diseases, including Retinoblastoma with Intracranial Spread Along Cranial Nerve (1200), Sjögren's Syndrome (1056), Bilateral Osgood-Schlatter Disease with Chronic Inflammatory Arthritis (665), Neurofibromatosis Type 1 (NF1) with Optic Glioma (944), Tuberous Sclerosis (664), and Brain Tumors (1048).

This dataset can be applied to a wide range of computer vision as well as image processing, image recognition and classification, segmentation, deep segmentation, and deep learning in vision. Its diverse conditions make it a valuable resource for advancing medical research and developing innovative diagnostic tools in healthcare.