Enhancing MRI-Based Brain Tumor Diagnosis with IDTNet: A Hybrid CNN Approach

Abstract

Early and accurate classification of brain tumors is of utmost significance to diagnosis and treatment planning. In this study, we introduce a deep learning-based approach to classify brain MRI images with Inception DenseNet with Transitions Network (IDTNet), a novel hybrid convolutional neural network that combines the strengths of Inception modules, DenseNet connectivity, and transition layers with skip connections. The model makes use of a dataset that contains 7022 MRI images that were obtained from figshare, the SARTAJ dataset, and Br35H, among other sites, Evaluated IDTNet against three CNN baseline architectures derived from VGG16, DenseNet121, and InceptionV1. Our proposed model has trained from scratch with aggressive data augmentation applied to improve generalization. IDTNet achieves a best accuracy of ∼98%, performing better than the baselines and achieving 100% recall on the meningioma class. These findings underscore the effectiveness of hybrid architectures in tackling complex imaging tasks.