Optimizing Deep Learning Based Approach for Brain Tumor Segmentation in Magnetic Resonance Imaging(MRI) Scans

Several medical imaging domains, including identification, segmentation, classification, and registration of imaging data, employ deep learning techniques. Brain tumors are one of the major causes of cancer-related deaths globally. The high death rate associated with brain cancer and the large population affected by the disease underscore the need for rapid and affordable brain cancer treatment. The early identification and diagnosis of brain tumors is critical to the prognosis for recovery from therapy. Magnetic resonance imaging (MRI) is a frequently used imaging modality for the diagnosis, treatment, and recovery of brain malignancies. However, it takes a lot of effort and specialized knowledge to manually identify brain cancers from a huge number of MRI scans. Our study aims to develop a deep learning system capable of identifying and categorizing brain cancers. While convolution neural networks (CNNs) are used to classify brain cancers, the U-Net model is utilized to segment MRI images. Performance measures including recall, accuracy, and precision are utilized to assess how effective this strategy is. The accuracy of the proposed CNN classifier was about 98% for both training as well as validation sets. Using DNN techniques, the following system seeks to provide accurate tumor segmentation and classification.