TOWARDS SMART NEURO DIAGNOSIS: CLOUD AND IOMT BASED BRAIN TUMOR DETECTION USING HIERARCHICAL DEEP LEARNING

Abstract

Brain tumors may be the key source of psychiatric complications such as depression, panic attacks, etc. The most important point towards curing the tumor is that recognition of a tumor in the brain should be timely and fast. The use of Medical Image Processing is having a very significant role in assisting the man in diagnosing the various illnesses. An important element is the Brain Tumor classification which depends on the experience and the knowledge of the doctor. An intelligent brain tumor detection and classification system is necessary to aid the doctors and physicians. The uniqueness of the research is that Brain tumors were separated into Glioma, Meningioma, and Pituitary using a hierarchical deep learning technique. The classification of the tumor and its diagnosis is very crucial in the rapid and effective treatment and processing of medical images with the Convolutional Neural Network is bringing great results in this regard. Convolutional Neural Network (CNN) exploits the pieces of the image as a way of training the information and categorizing the images into the classes that are later to be types of tumors. CNN is proposed to detect and classify brain tumors and create Hierarchical Deep Learning-Based brain tumor classifier. The suggested system categorizes the data entry into four categories that are named as Glioma, Meningioma, Pituitary and No-Tumor. The proposed model attained 96.54% Accuracy and Miss Rate is 3.46% better as compared to the earlier projects of segmentation and diagnosis of brain tumor. The suggested system will provide the clinical support in the field of medicine.