U-Net Model Based Classification on Brain Tumor in Magnetic Resonance Imaging (MRI) Multimodal

The U-Net model-based classification is intended to create a deep learning-based on automated system for segmenting of brain tumours. Users may submit MRI brain scan pictures and utilize the technology to separate growth from the brain tissue that is healthy. The system consists of a deep learning framework, a database for tracking patient information and segmentation results, and a UI-equipped web application. The U-Net architecture deep learning model, which has been successfully utilized for medical picture segmentation tasks, is the one employed in the system. A dataset of brain MRI images with matching manual segmentations serves as the model's training data. It is pre-processed using Z Score and Min Max Scaler to get it ready for training. Users of the 3D visualization may also interact with it by rotating the picture of the brain to see the tumour from various perspectives. The system generates comprehensive reports for each patient, providing detailed information about their tumours. These reports include specifics on tumour size, location, and grade. Tumour grade is assessed based on the patient's age and the tumour's dimensions, while the tumour's position is determined by analysing the segmented tumour image to identify its leftmost, rightmost, bottommost, and topmost boundaries. In conclusion, the automated brain tumour segmentation system the study's creation holds the potential to help doctors diagnose brain tumours accurately and quickly. With more training data and deep learning model tinkering, the system might achieve even more promising outcomes.