Deep Learning for Brain Tumor Classification: CNN Models with Saliency Map Insights

This paper aims to classify brain tumors using Convolutional Neural Networks (CNNs) applied to MRI images. We employed four models: a custom CNN, VGG16, ResNet101, and a hybrid model combining VGG16 and ResNet101, with saliency maps used for model interpretability. The custom CNN achieved the highest accuracy of 98.01%, followed by VGG16 (96.33%), ResNet101 (90%), and the hybrid model (97.10%). Saliency maps provided insights into the features driving predictions, enhancing model transparency. Our findings demonstrate the effectiveness of CNNs for brain tumor classification, with the custom CNN performing the best, and suggest further research into model explainability, data augmentation, and hybrid techniques to improve accuracy and generalizability.