Brain MRI ND-5 Dataset

1.     Download the Dataset:

• The dataset is provided in a single zip file. Please download the zip file from the provided link.

2.     Unzip the File:

• After downloading, unzip the file. You will find two main folders:
  • Training: Contains images for training your model.
  • Testing: Contains images for testing and validating your model.

3.     Folder Structure:

• Inside both the Training and Testing folders, there are four subfolders corresponding to different classes:
  • Glioma: MRI images of patients with glioma tumors.
  • Meningioma: MRI images of patients with meningioma tumors.
  • Pituitary: MRI images of patients with pituitary tumors.
  • No-Tumor: MRI images of patients without any tumor.

Each subfolder contains labeled images of the respective category, structured for easy use in classification tasks.

4.     Usage Guidelines:

• Use the images in the Training folder for model training and the images in the Testing folder for evaluation.
• You can apply any pre-processing, augmentation, or normalization techniques suitable for MRI image analysis.
• Ensure that your model outputs correspond to the four categories: glioma, meningioma, pituitary, and no-tumor.

5.     Suggested Workflow:

• Split the dataset as necessary if you need additional validation sets.
• Utilize machine learning or deep learning frameworks such as TensorFlow, PyTorch, or Keras for classification tasks.
• Implement multi-class classification techniques to distinguish between the four categories.

This dataset was created as part of our work, where we typically requires a large amount of diverse data to achieve optimal performance. To meet these demands, we combined data from five publicly available brain tumor MRI datasets, resulting in a more robust and comprehensive dataset. By merging these datasets, we have increased the diversity and quantity of the data, which enhances the robustness of models trained on this dataset for tasks like brain tumor classification and detection. Below are the original datasets that were combined to create this new dataset:

1.     Brain Tumor Classification (MRI) by Sartaj Bhuvaji

2.     Brain Tumor Dataset by Deniz Kavi

3.     HR Pathology Dataset by Kolarluni

4.     Tumour Classification Images by Lakshan Cooray

5.     Brain Cancer Detection (MRI Images) by Hamza Habib

We sincerely acknowledge the contributions of the authors of these datasets, whose work has laid the foundation for creating a more diverse and effective dataset. Source links are in documentation file. 

By combining these five datasets, we aim to provide a stronger, more generalized dataset for brain tumor classification and detection.
If you use this dataset, please be sure to cite both this work and the original datasets listed above.

The dataset is currently under embargo and will become publicly available on December 31, 2025. Until then, access is restricted to protect the integrity of the research.