Optimizing Brain Tumor Recognition with Ensemble support Vector-based Local Coati Algorithm and CNN Feature Extraction

Abstract

Abstract Nowadays, brain tumor (BT) recognition has become a common phenomenon in the healthcare industry. In the medical system,BT identification and classification can take a significant part in the diagnostics and considerations of the patients. BT is characterized as an abnormal mass of tissue in which the cells proliferate unexpectedly with no control over cell proliferation. In recent years, improvements in machine learning (ML), particularly deep learning (DL) procedures, have shown significant potential for mechanizing and improving these undertakings by utilizing medical imaging information. Also, we examine the difficulties and probabilities in this field, including information shortage, model interpretability, and moral contemplations. To overcome these challenges Ensemble support Vector-based Local Coati (ESV-LC) Algorithm is employed to identify and classify the brain tumor disease in the patients. For optimal classification, the features need to be extracted and this can be achieved by employing the Convolutional Neural network (CNN). To accurately classify BT, Ensemble Support Vector Machine (ESVM) is involved, which enhances classification performance, and hyperparameter tuning is performed through Local Search Coati Optimization. The Brain Tumor Image Dataset and Figshare Brain Tumor dataset are utilized for BT classification and identification. The performance metrics like Accuracy, Precision, Sensitivity, Specificity, and F1-score are to be evaluated, where the accuracy achieves the value of 98.3%, sensitivity of 97.6%, precision of 97.7%, specificity of 98.1%, and F1-score of 96.7% respectively.