Automated Brain Tumor Localization and Detection from MRI Scans Using Deep Learning with Attention-Based MultiScale Feature Fusion

Abstract

fundamental challenge in Deep Learning, Medical Imaging, MRI Analysis. Existing approaches, including VGG16, ResNet50, and IncepV3, process input data at a single resolution and fail to capture patterns spanning multiple scales, resulting in a mean accuracy ceiling on benchmark datasets. We address this limitation by introducing BrainTumorNet, a hybrid deep learning framework that integrates three parallel convolutional streams (kernel sizes 3, 7, and 13) with bidirectional LSTM encoding and a gated attention fusion module. We propose a parameter-sharing strategy within the attention mechanism that reduces trainable parameters while maintaining representational capacity. We train and evaluate our framework on BraTS 2023, Figshare using stratified 10-fold cross-validation. Our method achieves a mean dice score (%) of 96.3% on the primary benchmark, surpassing the nearest baseline by 3.5 percentage points (p < 0.001, Cohen's d = 1.42). We further demonstrate a 32.8% reduction in computational cost relative to comparable hybrid architectures and convergence within 155 epochs on all benchmark datasets. Keywords Brain tumour; CNN; deep learning; MRI; segmentation; transfer learning