Brain Tumor Classification Based on Deep Learning Techniques: An Extensive Study

Out of all diseases, brain tumors are the most deadly. The likelihood of patient surviving a intracranial tumor can be improved by early detection. Manual study of the MRI images by radiologist, which is extremely time consuming, can be replaced by computer-aided diagnosis. The CAD techniques include several stages such as tumor detection, region of interest segmentation, and tumor classification, for identifying of tumors from the MRI scans. The evolution of deep learning methods has greatly simplified it to identify brain tumors. This article offers a thorough analysis of the major accomplishments of the strategies listed and focuses on examining the most recent developments in brain tumor diagnosis using deep learning techniques. Additionally, it provides a comparison of three classification models—ResNet50, CNN, and VGG16, in which ResNet50 outshined the other two.