<p>Early detection of brain tumors is critical for effective treatment. Recent studies have demonstrated the effectiveness of Convolutional Neural Networks (CNNs) in medical image classification, particularly in detecting brain tumors. This research explores the use of the EfficientNet architecture, a state-of-the-art CNN model, for brain tumor detection. The goal is to create a highly accurate and efficient system leveraging EfficientNet’s capabilities. The methodology involves training the EfficientNet model on a diverse dataset of brain images, including both tumor and non-tumor samples. This dataset is carefully curated to ensure a balanced representation of tumor types and imaging modalities. During training, the model extracts high-level features from the images and classifies them as tumor or non-tumor. To optimize performance, techniques like data augmentation (rotation, scaling, flipping) and transfer learning with pre-trained weights are employed. The model is evaluated using an independent test dataset, with metrics such as accuracy, sensitivity, specificity, precision, and F1-score assessing its effectiveness. To enhance model transparency and build trust, the Grad-CAM technique is employed to visually explain predictions, enabling meteorologists to identify the specific image regions that contributed to the model’s decisions. A comparative analysis with other state-of-the-art methods reveals the EfficientNet model’s superior performance, achieving a remarkable accuracy of 98.83%. These results highlight the potential of CNN EfficientNet in improving diagnostic accuracy and efficiency in brain tumor detection, offering valuable insights for medical image analysis</p>