An efficient approach for binary classification in brain tumor detection using convolutional neural network

Abstract

Brain tumor detection using Convolutional Neural Network (CNN) models with binary classification has significantly improved the reliability of medical imaging through Deep Learning. The purpose of this research is to develop a modified CNN model by altering the different layers and weight values of each node to attain similar performance statistics to widely accepted CNN models while maintaining runtime efficiency. The proposed CNN model incorporates binary cross entropy to analyze the training data and accurately identifies whether or not a certain structured magnetic resonance imaging (sMRI) picture contains a tumor. In comparison to existing pre-trained CNN models, this study aims to contribute to the computer-aided diagnostic (CAD) system by implementing the proposed model with a simplified time complexity. The model achieved an overall classification accuracy of 96.7% after extensive tweaking of the proprietary CNN architecture. The suggested system’s performance is also compared with other existing systems, and the study demonstrates that it performs on par with most of them.