Brain Tumor Detection Using Deep Learning

Abstract

Brain tumors, arising from abnormal cell growth in the brain, can be life-threatening without accurate detection and treatment planning. Early detection and accurate classification of brain tumors are crucial for effective treatment planning and improved patient outcomes. Traditional methods like manual MRI examination and rule-based algorithms often lack precision, resulting in inconsistent detection and classification of brain tumors. In this work, the brain tumor detection and classification system is proposed to resolve these problems. This model uses CNN to accurately detect and classify brain tumors into types such as glioma, meningioma, and pituitary tumors. Using a dataset of 7,022 brain MRI images, the model utilizes advanced CNN architectures, including ResNet-50, DenseNet, and EfficientNetB1 for accurate detection and classification of tumors. However, these models tend to suffer from overfitting issues, affecting their generalization capability. To address this, we also employ InceptionV3, which demonstrates superior accuracy and robustness in detecting and classifying brain tumors. Therefore, patients will receive an accurate diagnosis, and the necessary treatment plan will be developed.