Brain Tumour Detection Using Deep Learning Techniques

Brain tumor detection is one of the most crucial problems in medical imaging solving which machine learning and, recently, deep learning techniques have demonstrated promising outcomes. This article explores deep learning algorithms for brain tumor detection in the proposed work. For this, the pre-trained architectures such as CNNs EfficientNetB1, VGG16, ResNet50, Inception, and MobileNetV2 are employed. The well-curated dataset of approximately 3400 MRI images of four tumor classes is used to train and evaluate the model using the performance metrics precision, accuracy, F1 score, and recall. EfficientNetB1, which is characterized by the efficient scaling properties, shows results-based on high accuracy and efficiency of 97.72%, and an ensemble and combination with SCNN demonstrates results of 98% accuracy, which improves the work overall. The study also delves into interpretability, elucidating the fusion of features extracted by the ensemble model. Overall, the research highlights the efficacy of ensemble learning in medical diagnosis, showcasing the potential of combining DL techniques to advance healthcare applications and increase the precision of brain tumor detection.