Attention-Infused U-Net Framework for Robust Brain Tumor Segmentation and Classification

Brain tumor segmentation and classification are essential in medical image analysis, enabling accurate diagnosis and effective treatment planning. Manual analysis of MRI scans is time-intensive, prone to human error, and requires specialized expertise, necessitating the development of automated techniques for improved detection accuracy and patient outcomes. This study proposes a brain tumor classification model implemented in Python, integrating MultiRes U-Net with an Adaptive Inception Net. The model operates with a population size of 15 over 65 iterations, achieving high accuracy (98.4282%), sensitivity (98.4884%), specificity (98.4172%), and precision (97.4177%). Comparative evaluation against existing deep learning models demonstrates its superior performance, as evidenced by the highest F1-score (97.4384%) and Matthews Correlation Coefficient (90.8171%), along with the lowest False Discovery Rate (3.2514%). These findings emphasize the effectiveness and dependability of the MultiRes U-Net combined with the Adaptive Inception Net model in brain tumor segmentation and classification, contributing to enhanced clinical decision-making and patient outcomes.