Convolutional Block Attention Module-based Deep Learning Model for MRI Brain Tumor Identification (ResNet-CBAM)

Uncontrolled and rapid cell growth in the body, leading to the formation of tumors, poses significant health risks by affecting surrounding healthy cells and can prove fatal if not treated at an early stage. Despite numerous efforts and promising results, accurately predicting and classifying tumors remains challenging. This research aims to enhance early detection and tumor growth recognition from MRI scans, facilitating prompt treatment initiation. Specifically, this study focuses on detecting lower-grade gliomas in the human brain using a dataset of MRI images from 110 patients. A deep learning approach utilizing the ResNet-50 model integrated with Convolutional Block Attention Module (CBAM) is employed for tumor detection. Following preprocessing and training with segmented tumor masks, the model achieved significant accuracy levels. Brain tumors, a severe type of malignant illness, can drastically reduce life expectancy if not detected early. They vary in type, grade, and location, influencing treatment strategies. MRI, though specialized, time-consuming, and costly, is essential for brain tumor diagnosis. Enhancing the accuracy and efficacy of MRIbased tumor identification and categorization is the goal of this research endeavor. ResNet-50, a pre-trained model, demonstrated high reliability for detecting and classifying medical images, achieving a notable accuracy rate. These improved accuracy rates are crucial for early tumor diagnosis, helping prevent severe physical consequences such as paralysis and disability.