An Empirical Review on Brain Tumor Classification Approaches

Cancer or malignant growth is the significant medical issue that shatters the world due to higher mortality rate. Cancer is regarded as the second most deadliest disease after cardiovascular disease as it kills one sixth of people in the world. There are different sorts of cancers such as colon cancer, lung cancer, breast cancer, blood cancer and brain tumor. The Brain tumor prevails as the most destructive cancer due to its low survival rate, assorted characteristics and combative nature. The brain is the crucial part of the human body as it controls the whole activity such as breathing, muscle movement and inducing the senses with the help of tissues and neural cells. Each and every cell has their own abilities; a few cells develop with their own usefulness, and some lose their ability, oppose, and develop distortions. These mass assortments of unusual cells from the tissue are called cancer. Harmful brain tumors are uncontrolled and unnatural development of synapses (brain cells). It is quite possibly the most dangerous and deadly malignant growth, which forces the requirement for programmed recognition techniques. There are so many methods existing in the literature to facilitate effective brain tumor classification methods. Hence, in this survey, we analyze 30 literature works concentrating on classification problems associated with the brain tumor using MRI, bringing into light the various shortcomings of the existing methodologies for classification problems. Here, the analysis of various methods will be facilitated based on several factors, such as performance metrics, datasets used etc. On the other hand, the analysis of the methods with respect to the merits and demerits of the methods are presented. Finally, the paper elaborates the prospective future research directions and provocations in obtaining the better classification accuracy for the MRI brain tumor categorization.