Deep Learning Based Multiclass Tumor Identification and Classification Using Fusion of CNN Models

The rising incidents of brain tumors pose significant challenges in diagnosis and treatment. Tumorous brain cancers present unique difficulties due to their critical locations and functional implications. Magnetic Resonance Imaging (MRI) offers safe and precise imaging, free from the risks associated with radiation-based methods. Deep learning, particularly Convolutional Neural Networks (CNN), has emerged as a valuable tool in identifying brain tumors in MRI scans, enabling more accurate diagnoses by fusing two well-known DL models. This study explores a fusion strategy combining AlexNet and VGG16 models, outperforming individual models trained with different learning rates. While VGG16 and AlexNet achieved accuracies of 79% and 77% respectively. Our fusion approach potentially achieved an accuracy of 81%, underscoring its efficacy in enhancing brain tumor classification accuracy. This study helps medical professionals to accurately and timely classify tumors for better outcomes for patients.