Abstract

The diagnosis of brain tumors is a major issue in medical imaging, having the potential to have a large influence on patient outcomes. This abstract provides an overview of the obstacles, most recent breakthroughs, and ongoing research in brain tumor detection. Early diagnosis requires medical imaging modalities such as computed tomography (CT) and magnetic resonance imaging (MRI). Previously, radiologists were evaluated manually, but with the emergence of deep learning algorithms, particularly Convolutional Neural Networks (CNNs), the field has experienced a revolution. When paired with large and diverse datasets, these AI-based algorithms have demonstrated promising results in automating the diagnosis of brain tumors, boosting accuracy, and reducing the workload for medical workers. However, there are still challenges, such as the need for larger datasets and the interpretability of AI models, even though CNN attained a very outstanding accuracy of 97.87% in our study. The fundamental purpose of this research is to distinguish between normal and abnormal pixels using statistical and texture-based criteria.