Brain tumor diagnosis based on convolutional neural network improved by a new version of political optimizer

Abstract

New strategies for the prevention, diagnosis, and treatment of brain tumors have emerged as a result of advances in medical science. A brain tumor is recently one of the best cases for the focus of researchers in a variety of scientific domains, including medicine, medical engineering, and software engineering, as a result of its aggressive nature. Accurate tumor area recognition in MRI scans has a multitude of purposes, including determining the kind of illness, analyzing treatment and disease progression, and creating therapeutic measures in radiation approaches. Currently, this is made by humans. Considerable study has been done to introduce computer processing in medical sciences, particularly in the processing of medical MRI images, as science has progressed. MRIs of the intracranial cavity offers a comprehensive view of the brain in most cases. A doctor examines this visual representation in order to identify a brain tumor. This method of diagnostics, on the other hand, is incompatible with pinpointing the specific position and size of the tumor. To get a more accurate diagnosis, in this study, we designed an automatic methodology based on deep learning and optimization for diagnosing brain tumors. The method uses an improved version of the political optimizer for the optimal arrangement of the convolutional neural network (CNN) which is done by optimal selection of the CNN hyperparameters. The designed model has been then implemented on the Figshare dataset and its outcomes are verified and validated with some other processes to indicate its efficiency.