Detection and Classification of Brain Tumour Using EfficientNet and Transfer Learning Techniques

Medical imaging plays a crucial role in the diagnosis and categorization of brain tumours. Recently, deep learning methods, specifically Convolutional Neural Networks (CNNs), were effectively utilised to this issue. In this research, we propose employing a Convolutional Neural Network (CNN) architecture based on EfficientNet for the identification and classification of brain tumours. Preprocessing, feature extraction, and classification are the three key components of our proposed model. During the initial stages of preparation, we apply skull stripping and intensity normalization techniques to remove unwanted noise and enhance image contrast. Next, we employ transfer learning with the EfficientNet architecture to extract relevant features from the preprocessed images. Finally, we use a fully connected layer with a softmax activation function to classify the extracted features into four categories such as healthy brain, glioma, meningioma, and pituitary tumor. Brain Tumour Segmentation (BraTS) is a dataset used for research into brain tumours, and we use it in order to evaluate our suggested approach, which contains 3D MRI scans of brain tumors. The experimental results demonstrate that our proposed model achieves high accuracy, sensitivity, and specificity for tumor detection and classification. Specifically, our model achieves an overall accuracy of 97.2 %, sensitivity of 96.8%, and specificity of 97.4%. These results outperform numerous state of the art methods, demonstrating the effectiveness of our proposed CNN architecture for brain tumor detection and classification. Overall, our study presents a promising approach for improving the accuracy and efficiency of brain tumor diagnosis through deep learning techniques.