Main Article Content
Classifying brain tumors using machine learning techniques have become an essential due to its importance in people life. The correct and fast diagnosis are the keys to reduce the percentage of deaths that have raised recently to significant numbers. The available techniques such as CT scan and MRI imaging are widely used nowadays and the latter is more common as it provides high resolution images from different angles for brain tissues. Determining the right type of brain tumor manually requires an expert who has a good knowledge in brain diseases. Also, it is time consuming and tedious for a lot of images. Moreover, human errors are possible and consequently false detection may cause a wrong procedure and treatment. Therefore, the scientists and researchers introduced different approaches for classifying tumor types automatically and efficiently without needing to human knowledge. This paper reviews these approaches, which includes traditional machine learning algorithms (MLs). These algorithms can be divided into two main parts, supervised and unsupervised. The most algorithms that are used and achieved high accuracy are SVM, KNN, and ANN. On the hand, today by enlarging the available data in this area and developing new ANN-based techniques, called deep learning, the performance of brain tumor classification is boosted. This family of techniques which can be used for both feature extraction and classification are also reviewed in this paper.