3D MRI Image Segmentation Using Ensemble Models

Brain tumors are among the deadliest diseases worldwide, with MRI and CT scans being the traditional methods for diagnosis. However, these techniques are time- consuming and prone to human error, potentially delaying crucial treatment decisions. To enhance diagnostic accuracy and efficiency, deep learning techniques have been increasingly integrated into medical imaging. Over the years, deep learning has shown exceptional performance in analyzing complex medical images, providing reliable support to healthcare professionals and facilitating the early detection and treatment of brain tumors. The analysis of 3D brain MRI volumes is a time-intensive task for doctors. Deep learning algorithms assist in quickly identifying areas of interest, allowing doctors to finalize results more efficiently. Brain tumor segmentation is a process that separates the cancerous part of the brain from healthy tissue and is classified into four stages: preprocessing, segmentation, optimization, and feature extraction. MRI-based segmentation has gained importance due to its reliability, safety, and high resolution. This project proposes an automatic segmentation method based on convolutional neural networks (CNNs), which uses multiple layers for feature extraction and tumor segmentation