Enhanced Brain Tumor Image Classification Using Convolutional Neural Network with Attention Mechanism

Detecting brain tumors early is crucial for precise diagnosis and the development of effective treatment strategies, given the severity of the condition involving the uncontrolled growth of abnormal cell clusters in the brain.This research introduces an innovative Convolutional Neural Network (CNN) model augmented with an attention mechanism for the classification of brain tumor images, utilizing a comprehensive dataset of 3,000 Magnetic Resonance Imaging (MRI) scans sourced from Kaggle. Rigorous preprocessing procedures, encompassing image loading, color format conversion, resizing, and array conversion, ensured dataset standardization. The model architecture, featuring convolutional layers with increasing filter and kernel sizes, incorporated a pivotal Attention layer, enhancing feature representation through self-attention. In testing, the model exhibited a remarkable accuracy of 97.5% and a minimal test loss of 0.0975. Visual aids, including attention maps and overlays, were employed to elucidate the decision-making process, fostering model interpretability. These outcomes underscore the model's efficacy in brain tumor image classification, indicating its potential for reliable deployment in real-world applications. The accentuation of attention mechanisms and interpretability not only bolsters model trustworthiness but also furnishes valuable insights for medical professionals