*This dataset supports online use.Click here to jump.

BRISC 2025 is a magnetic resonance imaging (MRI) dataset for brain tumor segmentation and classification, which was released in 2025 by Iran University of Science and Technology, Shahrood University of Technology, University of Essex and other institutions. The related paper results are "BRISC: Annotated Dataset for Brain Tumor Segmentation and Classification with Swin-HAFNet".

The dataset contains approximately 6,000 T1-weighted MRI images, covering four categories: glioma, meningioma, pituitary tumor, and no tumor. All samples are accompanied by pixel-level segmentation masks reviewed by medical experts to ensure accurate annotation. The images cover the three major anatomical planes: axial, coronal, and sagittal. The images are stratified into 5,000 training samples and 1,000 test samples to support robust model training and evaluation.

This dataset supports two types of tasks: classification, which involves identifying multiple tumor types based on MRI images; and segmentation, which involves detecting tumor regions at the pixel level using paired MRI images and masks. Its clear organization and strict alignment of image and mask file names facilitate its direct application in the training and validation of deep learning models.