Analysis of Brain MR Images by using Transfer Learning Algorithms with Deep Learning Approach

This research paper analyses brain tumor MRI segmentation and classification using modified convolutional neural networks (CNN) and transfer learning with VGG16 techniques. These advanced methods improve detection, classification, and accuracy over conventional techniques, which typically involve feature extraction, pre-processing, image acquisition, and classification. MRI scans are utilized to diagnose various types of brain tumors, focusing on tumor type recognition and position identification. The study shows that the modified CNN and VGG16 techniques enhance accuracy, precision, recall, F1-score, macro average, and weighted average metrics compared to conventional methods. Fine-tuning the network algorithms’ hyper-parameters further increases precision. Specifically, the proposed methods achieve an average accuracy of 99.50% and 99.60% for weighted and macro averages, respectively, significantly outperforming traditional techniques.