A Brief Survey on Brain Tumor Identification Models Using Deep Neural Networks

Abstract

A brain tumour occurs when abnormal cells increase in the brain. Non-cancerous (Benign) brain tumours include astrocytoma, glioblastoma, and oligodendroglioma. Cancerous (Malignant) brain tumours include craniopharyngioma and ependymoma. The survival rate of a patient with a high risk of developing a brain tumour is difficult to estimate due to the limited occurrence of brain tumours and their assortment. About 15 out of every 100 persons with brain cancer will have a chance of survival for ten or more years after being diagnosed, according to the UK study. The sort of tumour, how abnormal the cells are, and where it is in the brain all influence how the patient will be treated for brain tumour. Advances in machine learning are utilised to accurately diagnose the brain tumour by analysing Magnetic Resonance Imaging (MRI) pictures. The classification of brain tumours has a significant impact on clinical diagnosis and on providing the most appropriate treatment. The current practise of making a diagnosis and classifying brain tumours still depends on microscopic study of biopsy samples. At now, the process is lengthy, complex, and error-prone. While these drawbacks illustrate how important it is to have a completely automated classification process for many classifications of brain tumours, deep learning offers an even better alternative. This paper presents a brief survey on several models for brain tumour detection using deep learning techniques.