A ResNet-Powered Approach for Brain Tumor Detection with Particle Swarm Optimization

Le résumé fourni par la source

Brain tumors cause significant distress, affect the brain or surrounding tissues, and cause severe damage. Timely and accurate brain tumor diagnosis is essential for effective treatment and improving patient outcomes. Recent advancements in deep tissue research, particularly in the field of neuroimaging, offer hope and potential. Magnetic resonance imaging (MRI) is still used to diagnose brain problems because it generates precise brain structure images non-invasively utilizing powerful magnetic fields and radio frequencies. In this research, we proposed a unique method for classifying brain tumors into three distinct classes: meningioma, pituitary, and glioma, using a Residual Neural Network (ResNet) architecture and Particle Swarm Optimisation (PSO) to maximize efficiency. Renowned for its deep learning capabilities, the ResNet50 architecture has been optimized to efficiently extract complex information from images of brain tumors. We have optimized the classification process by fine-tuning the network’s hyperparameters to maximize its discriminative capability by utilizing PSO. The combination of optimization and deep learning methods has produced remarkable outcomes, raising the bar for brain tumor classification accuracy. The results show that our method routinely achieves 99.3% test accuracy, outperforming previous techniques and creating a new standard for brain tumor classification.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.