The human brain is a crucial organ composed of billions of cells. When cells in the brain undergo abnormal and uncontrolled division, it can result in the formation of tumors, which can exert immense pressure on nerves and blood vessels, potentially causing irreversible harm to the body. An extensively employed method to identify brain tumors is Magnetic Resonance Imaging (MRI). The term IoMT represents the Internet of Medical Things, which refers to the combination of networked technology with medical devices and applications for remote monitoring and management of patient health. Advanced machine learning algorithms and image segmentation techniques can classify these tumors. This study utilized a Convolutional Neural Network framework, specifically ResNet-50, for classification of brain tumors into three categories: meningioma, glioma and pituitary tumor. By incorporating ResNet-50, the study was able to tackle the problem of vanishing gradient and facilitate training of deeper neural networks. The MRI image dataset is preprocessed and augmented to improve image quality. The CNN ResNet-50 framework ensures effective super-resolution of brain images, improving tumor classification accuracy, especially for early cancer nodules. Performance is evaluated based on training and testing accuracy.