Collective Intelligence: Harnessing Multiple Models to Enhance Brain Tumor Detection and Classification Utilizing Deep Learning Techniques

Brain tumors arise from abnormal cell growth near or within the brain due to genetic changes. They pose a significant health risk, particularly in children, ranking as the second most common childhood cancer in the United States. Existing detection methods, though necessary, are invasive, time-consuming, and prone to errors. Therefore, there is a pressing need for a more efficient and reliable approach to improve diagnostic accuracy and speed. This study explores collective intelligence by employing multiple deep learning models using transfer learning techniques for brain tumor detection and classification. It combines InceptionV3, DenseNet201, and ResNet50 to create an ensemble that enhances performance. Based on 1,311 high-resolution brain MRI scans covering four types of brain tumors, the research applies data preprocessing and augmentation to improve dataset quality and diversity. The ensemble method addresses the limitations of single models, increasing the reliability and robustness of detection and classification. The ensemble model achieved superior performance, with an accuracy of 95.6%, an F1-score of 94.4%, and a precision of 94.9%, underscoring the effectiveness of ensemble learning in enhancing brain tumor diagnostic capabilities.