<b>BRISC: Annotated Dataset for Brain Tumor Segmentation and Classification</b>

<p dir="ltr"><b>BRISC 2025 — Brain Tumor MRI Dataset</b></p><p dir="ltr">BRISC (BRain tumor Image Segmentation & Classification) — a curated, expert-annotated T1 MRI dataset for multi-class brain tumor classification and pixel-wise segmentation.</p><p dir="ltr"><b>ArXiv preprint (Fateh et al., 2025): https://arxiv.org/abs/2506.14318</b></p><h2>�� Overview</h2><p dir="ltr">BRISC is designed to address common shortcomings in existing public brain MRI collections (e.g., class imbalance, limited tumor types, and annotation inconsistency). It provides high-quality, physician-validated pixel-level masks and a balanced multi-class classification split, suitable for benchmarking segmentation and classification algorithms as well as multi-task learning research.</p><p dir="ltr">Highlights<br>- 6,000 T1-weighted MRI slices (5,000 train / 1,000 test)<br>- Four classes: Glioma, Meningioma, Pituitary Tumor, No Tumor<br>- Pixel-wise segmentation masks reviewed by radiologists<br>- Slices from three anatomical planes: Axial, Coronal, Sagittal<br>- Clean, stratified train/test splits and aligned image–mask filenames</p><h2>�� Dataset structure</h2><p dir="ltr"><b>brisc2025/</b><br>├─ classification_task/<br>│ ├─ train/<br>│ │ ├─ glioma/<br>│ │ │ ├─ brisc2025_train_00001_gl_ax_t1.jpg<br>│ │ │ └─ ...<br>│ │ ├─ meningioma/<br>│ │ ├─ pituitary/<br>│ │ └─ no_tumor/<br>│ └─ test/<br>│ ├─ glioma/<br>│ │ ├─ brisc2025_test_00001_gl_ax_t1.jpg<br>│ │ └─ ...<br>│ ├─ meningioma/<br>│ ├─ pituitary/<br>│ └─ no_tumor/<br>├─ segmentation_task/<br>│ ├─ train/<br>│ │ ├─ images/<br>│ │ │ ├─ brisc2025_train_00001_gl_ax_t1.jpg<br>│ │ │ └─ ...<br>│ │ └─ masks/<br>│ │ ├─ brisc2025_train_00001_gl_ax_t1.png<br>│ │ └─ ...<br>│ └─ test/<br>│ ├─ images/<br>│ │ ├─ brisc2025_test_00001_gl_ax_t1.jpg<br>│ │ └─ ...<br>│ └─ masks/<br>│ ├─ brisc2025_test_00001_gl_ax_t1.png<br>│ └─ ...<br>├─ manifest.json<br>├─ manifest.csv<br>├─ manifest.json.sha256<br>├─ manifest.csv.sha256<br>└─ README.md<br><br></p><p dir="ltr"><b>Notes:</b><br>- Classification folders contain image-level labels suitable for standard image classification pipelines.<br>- Segmentation folders contain paired MRI images/ and corresponding binary masks/.<br>- Image and mask filenames are identical except for file extension (images: .jpg, masks: .png).<br>- All images are T1-weighted slices.</p><h2>�� Dataset statistics</h2><p><br></p><p dir="ltr">- Total samples: 6,000 (5,000 train / 1,000 test)<br>- Classes: 4 (balanced distribution across train/test)<br>- Planes: Axial / Coronal / Sagittal (balanced representation)<br>- Imaging modality: T1-weighted MRI<br>- Annotation quality: Reviewed and corrected by medical experts</p><h2>�� Citation</h2><p dir="ltr">If you use BRISC in your work, please cite:</p><p dir="ltr"><code>@article{fateh2025brisc,</code><br><code>title={Brisc: Annotated dataset for brain tumor segmentation and classification with swin-hafnet},</code><br><code>author={Fateh, Amirreza and Rezvani, Yasin and Moayedi, Sara and Rezvani, Sadjad and Fateh, Fatemeh and Fateh, Mansoor and Abolghasemi, Vahid},</code><br><code>journal={arXiv preprint arXiv:2506.14318},</code><br><code>year={2025}</code><br><code>}</code></p><h2>�� Acknowledgments</h2><p dir="ltr">Thanks to the collaborating radiologists and physicians for expert annotation and review.</p><h2>�� References & inspirations</h2><p dir="ltr">This dataset drew design and organizational inspiration from widely used brain tumor imaging datasets (e.g., BraTS, Figshare datasets, Kaggle collections). See the project paper for full details and evaluation results.</p>