A Robust Approach for Brain Tumor Detection using Transfer Learning

Brain tumor is a significant health concern that requires an early diagnosis for improved treatment outcomes. Artificial Intelligence (AI) can assist to reduce the need for invasive biopsies by analyzing medical imaging data to successfully detect brain tumor. Transfer learning is an effective technique in the field of deep learning, it has developed into a promising approach for improving model performance. In this study, transfer learning model VGG16, Inception V3 and EfficientNet B2 used for brain tumor detection based on their texture, location, and shape. The Kaggle (Brain Tumor Classification - MRI) dataset was utilized for training and testing the model. The outcome of the proposed model state-of-the-art methods on result demonstrates that it is more accurate and could assist doctors make better decisions.