Transfer Learning Using EfficientNet for Brain Tumor Classification from MRI Images

A benign or malignant brain tumour is a development of cells within the brain or skull that is abnormal. A primary tumour is one that grows directly from the brain's tissue, while a secondary tumour is one that spreads from another part of the body to the brain (metastasis). Depending on the tumor's kind, size, and location, there are a variety of treatment options. Radiologist use MRI (magnetic resonance images) to classify them. Due to the complexity of brain tumours and their qualities, a manual examination might be error-prone. To aid physicians throughout the globe, we've proposed a solution that uses Deep Learning Algorithms like ConvolutionNeural Networks and TransferLearning (TL). In this paper we have classified MRI into four types (glioma, no tumor, meningioma, pituitary). We have trained our architecture using EfficientNet. From According to our findings, transfer learning works well when the dataset is limited. An accuracy of 99 percent is achieved by the suggested approach.