A Hybrid Model for Brain Tumor Segmentation using VGG16 and ResNet50

Abstract

Cancers of the brain are among the worst illnesses a person may get. The course of medical therapy is mostly determined by the tumor's location and kind. Neuro specialists and radiologists must carefully review Magnetic Resonance Imaging (MRI) pictures in order to arrive at a definitive diagnosis of a malignancy. Treatment option mapping, disease progression monitoring, and image-based tumor segmentation are of utmost importance in medical imaging because they provide information vital for cancer analysis and diagnosis. Some have speculated that deep learning might be the key to better brain cancer diagnosis and treatment. With its state-of-the-art segmentation and detection capabilities, the segmentation strategy has significantly improved the removal of abnormal tumor regions from the brain. In this study, we published a VGG16 and ResNet50 hybrid model for MRI brain tumor segmentation. With the use of the ResNet50 algorithm and the VGG16, a Transfer Learning method, brain tumors may be detected in segmented pictures. The findings show that our suggested approach is more accurate and performs better than other models when compared side by side.