Inspiration ## Inspiration

Brain tumors require accurate diagnosis and careful analysis of MRI scans. We were inspired by the potential of Artificial Intelligence to assist medical professionals by enhancing MRI images and accurately identifying tumor regions.

What it does

Our project enhances brain MRI images using CLAHE and performs 3D brain tumor segmentation using a 3D U-Net deep learning model. It processes FLAIR, T1ce, and T2 MRI modalities and identifies tumor regions from 3D scans.

How we built it

We built the project using Python, TensorFlow/Keras, NumPy, NiBabel, scikit-image, Pandas, and Matplotlib. We used the BraTS dataset for training and testing, applied MRI preprocessing and enhancement, and trained a 3D U-Net model for tumor segmentation.

Challenges we ran into

The major challenges were handling large 3D MRI volumes, preprocessing medical images, managing memory and computational requirements, and achieving accurate segmentation of different tumor regions.

Accomplishments that we're proud of

We successfully developed a pipeline that combines MRI enhancement with AI-based 3D tumor segmentation. We are proud of integrating medical image processing and deep learning into a single system.

What we learned

We learned about MRI image processing, medical image formats, 3D deep learning, U-Net architecture, dataset preprocessing, model training, evaluation metrics, and the challenges involved in working with medical imaging data.

What's next for Brain Tumour MRI Enhancement + 3D Segmentation

In the future, we plan to improve segmentation accuracy, support more MRI modalities, optimize the model for faster processing, develop a user-friendly interface, and explore deployment as a clinical decision-support tool.