Brain Tumor Classification MRI Dataset (Glioma, Meningioma, Pituitary, and No Tumor)
Description
This dataset is designed for the development and evaluation of machine learning and deep learning models for the automated detection and classification of brain tumors from MRI scans. It is structured to support multi-class classification tasks, distinguishing between three common types of brain tumors and healthy (no tumor) scans. The dataset is organized into two primary subsets: Train and Test, facilitating standardized model training and performance benchmarking. Dataset Structure The data is organized into a hierarchical folder structure: - Train/: Contains images used for model training and validation. - Test/: Contains independent images used for final model evaluation. Each of these folders contains four sub-directories representing the target classes: - glioma_tumor: Images showing glial cell tumors. - meningioma_tumor: Images showing tumors arising from the meninges. - no_tumor: MRI scans of healthy brains or non-tumorous cases. - pituitary_tumor: Images showing tumors of the pituitary gland. Potential Applications - Training Convolutional Neural Networks (CNNs). - Benchmarking Transfer Learning models (e.g., ResNet, VGG, EfficientNet). - Developing medical imaging diagnostic tools. - Data augmentation research for medical imaging.