Detection of Brain Tumor using a Feature Optimization using Particle Swarm Optimization and Ensemble Classifier

The brain tumor is a damning disease after cardiovascular diseases across the globe. However, brain tumor detection's accurate and early-stage saves millions of lives worldwide. This research introduced an ensemble-based classifiers in tumor detection, which was created using the bagging approach. The primary classifier in the ensemble classifier is the support vector machine and random forest classifier. Furthermore, the proposed ensemble classifier uses particle swarm optimization to work with the feature optimization method. The process of feature optimization enhances the feature selection process for the classifier. The brain tumor images are captured by magnetic Resonance imaging (MRI). The MRI images are rich in texture features, and now discrete wavelet transform applies for feature extraction. The BRATS dataset is being used to evaluate the suggested classification technique, which was implemented in MATLAB software. Extreme learning (EL) and CNN are compared to two existing algorithms in the suggested algorithm. The study of the results indicates that the suggested algorithm outperformed existing algorithms by 2%.