BRISC: Annotated Dataset for Brain Tumor Segmentation and Classification

TL;DR

BRISC addresses the scarcity and bias of public brain tumor datasets by providing 6,000 expert-annotated, contrast-enhanced T1-weighted MRI scans covering glioma, meningioma, pituitary, and non-tumorous classes, with masks across axial, coronal, and sagittal planes. The dataset emphasizes balanced class distributions, multi-institutional diversity, rigorous annotation quality, and thorough preprocessing, delivering robust train/test splits (5,000/1,000) and a comprehensive benchmark for segmentation and classification. The authors benchmark multiple baselines and introduce the Swin-HAFNet transformer as an effective framework for both tasks, demonstrating strong performance and generalization potential. BRISC is publicly accessible via Kaggle, Figshare, and Zenodo, enabling reproducible research and cross-institutional evaluation in brain-tumor imaging.

Abstract

Accurate segmentation and classification of brain tumors from Magnetic Resonance Imaging (MRI) remain key challenges in medical image analysis, primarily due to the lack of high-quality, balanced, and diverse datasets with expert annotations. In this work, we address this gap by introducing BRISC, a dataset designed for brain tumor segmentation and classification tasks, featuring high-resolution segmentation masks. The dataset comprises 6,000 contrast-enhanced T1-weighted MRI scans, which were collated from multiple public datasets that lacked segmentation labels. Our primary contribution is the subsequent expert annotation of these images, performed by certified radiologists and physicians. It includes three major tumor types, namely glioma, meningioma, and pituitary, as well as non-tumorous cases. Each sample includes high-resolution labels and is categorized across axial, sagittal, and coronal imaging planes to facilitate robust model development and cross-view generalization. To demonstrate the utility of the dataset, we provide benchmark results for both tasks using standard deep learning models. The BRISC dataset is made publicly available. datasetlink: Kaggle (https://www.kaggle.com/datasets/briscdataset/brisc2025/), Figshare (https://doi.org/10.6084/m9.figshare.30533120), Zenodo (https://doi.org/10.5281/zenodo.17524350)

Figures (8)

• Figure 1: Samples of Glioma segmentation across different imaging planes
• Figure 2: Samples of Meningioma segmentation across different imaging planes
• Figure 3: Samples of Pituitary segmentation across different imaging planes
• Figure 4: Samples from the non-tumorous class across different imaging planes. The top row shows healthy brain scans without visible abnormalities, and the bottom row shows examples of non-tumorous lesions (e.g., cysts or abscesses).
• Figure 5: Samples of whole-region misannotations. The red area indicates regions that were initially marked as tumors but were identified by radiologist and physician as non-tumorous.