An Efficient Model for Brain Tumor Classification Through Transfer Learning Approaches

Brain tumor classification is a critical task in medical diagnostics, where early and accurate detection can significantly improve patient outcomes. This work proposes an efficient model for brain tumor classification using CNN and transfer learning techniques. Initially, a base model CNN applied to MRI images of human brains collected from Kaggle which is classified into four categories. To enhance classification performance, four transfer learning models-VGG19, ResNet50, InceptionV3, and EfficientNetB0-are utilized alongside CNN. Each transfer learning model is fine-tuned and optimized to extract relevant features, improving classification accuracy and robustness. Model performance is evaluated using accuracy. Experiments demonstrated that the proposed approach given high accuracy, when compared to traditional models.